August 2016 update —- This part is complete through early 1921, or through the Oscar Denton time as observer at Greenland Ranch. A close examination of the following few decades of the Greenland Ranch record is needed, but that work will come later.
January 2019 update —- I am adding charts of the daily maximum temperatures for many periods during the Denton and pre-Denton years. These help to show how the regional daily maximum temperatures behave, and are useful for revealing possibly problematic data.
Some parts updated in January 2021. I added the daily temperatures (in pdf files) at Greenland Ranch from 1911 to 1921.
—-
This is Part 2 (of 6 parts or more!) of my research regarding the early temperature records for Death Valley, and especially its record maximum temperature of 134F. Part 1 looks closely at Death Valley weather station histories and previous studies on the 134F record. Part 2 dives into the first ten to twenty years of record at Greenland Ranch cooperative weather station and examines the data reliability. It will be shown that these early records are fraught with errors, inconsistencies, and poor observing practices. Because of the relatively routine problematic nature of the temperature data, it allows the researcher to make some assumptions regarding the character and capabilities of the observer(s). And, it makes it easier to dismiss, or reject, records which appear problematic.
The Death Valley record maximum of 134F has stood for 100 years. It is practically “etched in stone” in the minds of most climatologists and people with an interest in weather. It is not easy to undo records such as this one. If I just throw out a few statements and statistics, such as: “Death Valley’s long-term climatological record and the surrounding station data do not support the 130-degree temperatures.” And, “The observer back then was ill-suited for the task…” Well, that alone is not going to change a lot of minds. A solid body of research to support these contentions is required!
In order to convincingly discredit the 130-plus temperature reports at Greenland Ranch in July, 1913, there are two crucial elements to realize: a) the observer (i.e., Oscar Denton) could not be relied upon to provide trustworthy data, and b) maximum temperatures from surrounding stations do not support the record maximums in Death Valley. Before the detailed comparisons of high temperatures in the region, it is prudent to first delve into the minefield that is the early Greenland Ranch temperature record.
Early Greenland Ranch Record Woes
The reports received from any weather station are only as good as, only as reliable as, only as useful as the quality and reliability and proper exposure of the instruments and equipment; further, the records and reports from any weather station are only as good as, only as reliable as, only as useful as the competence, commitment, concern and due diligence of the person or persons responsible for the observations. The weather instruments at Greenland Ranch were carefully and properly set up by U.S. Weather Bureau personnel in 1911, and were left in the hands of the superintendent of the Pacific Coast Borax Company. The superintendent in 1911 was Fred Corkhill, and apparently he had other workers with the company perform the weather observing tasks. (In the Climatological Data publications through 1914, “J.W. Corkhill” is shown as observer for Greenland Ranch. Additional research indicates that Corkhill set up the instrumentation at Greenland Ranch, and that USWB personnel likely did not visit the station until 1924. Corkhill was an observer at the USWB station in Candelaria, Nevada, during the late 1890s, and was familiar with the equipment and observing procedures.) Little is known about the background of the observers at Greenland Ranch. (However Part 6 here provides some research on Oscar Denton.) Did they receive extra pay for taking the observations once a day? Were they thoroughly trained regarding the instrumentation and the proper procedure for making the observations? Did they fully appreciate the importance of providing consistent, accurate temperature and precipitation records? Did they welcome the weather observing job, or was the daily routine of taking the observations a nuisance or a “drag”? (New research on Greenland Ranch and Oscar Denton is in Part 6.)
The Greenland Ranch weather observer was tasked to read and reset the maximum and minimum thermometers in the wooden shelter just once per day, every day, at or around the official observation time of 5 p.m. It takes just a minute or so to carefully check the thermometers and to reset them. Then, of course, the observer had to enter the daily maximum temperature, the daily minimum temperature, and the “set max” temperature onto the monthly form. The “set max” temperature is the temperature shown by the maximum thermometer after it is spun a time or two. After a proper reset, it should show the current air temperature, and it should be very close (within a degree or so) of the temperature indicated by the minimum thermometer. Both thermometers should be in a horizontal position before the observer closes the shelter door. The daily temperature readings should reflect the highest and lowest shelter temperatures attained during the previous 24-hour period, since the previous day’s observation at 5 p.m. If there has been any rain since the previous observation, the observer is also tasked to determine the amount that has collected in the rain gage. Rain which enters the catchment portion of the standard 8-inch gage is funneled into a narrower metal tube, which allows readings to the nearest “hundredth of an inch” with a long, graduated, measuring stick.
This section scrutinizes the Greenland Ranch record from June, 1911, to 1921. Denton’s final month as observer was February, 1921, or thereabouts. His successor, Victor Ceballos, may have been filling in for Denton a little prior to then. There is quite a bit of comparison, detail and discussion, all of which is designed to provide a better understanding of observing at cooperative weather stations and of temperature patterns and behavior in the Death Valley region. One thing to keep in mind is that, when comparisons are made between Death Valley stations and surrounding stations, it is generally assumed, unless otherwise stated, that the data from the non-Death Valley stations are trustworthy, and are legitimate. The climate researcher should never blindly assume that the published weather data are good! Mistakes can and do occur in a variety of ways. It is the job of the climatologist to separate the good data from that which might be suspect.
This detailed review will show that a very large portion of the early Greenland Ranch climate record is riddled with errors, inconsistencies, suspect temperatures, curious data sets, and instrument malfunctions. The problematic data are due almost entirely to observer incompetence, carelessness, ignorance and/or laziness…and/or perhaps due in part to observer “over-zealousness.” If an observer does not know how to, or does not care to, provide a solid and reliable temperature record, then it is not realistic to expect consistently trustworthy data.
Forms for Independence are provided, also. The Independence weather station observer was the U.S. Weather Bureau, and the published data should be trustworthy and of excellent quality. Though Independence is about 75 miles west-northwest of Furnace Creek and Greenland Ranch, and about 4000 feet higher in elevation, the changes in temperature from day-to-day and week-to-week are comparable, especially during the warm half of the year. Both of the Greenland Ranch and Independence stations are near the bottom of their respective north-south and deep basins. Owens Valley and Death Valley are usually under the influence of the same air mass. If maximum temperatures are trending in opposite directions between the two stations for a period of time, then that may well mark a period of questionable reports from one of the stations.
New Section added to Part 2! I am compiling all of the Greenland Ranch daily max and min temperature (and precipitation) figures.
Well, at least all of them through 1921, for now. The entries in the table below are those original figures as they were entered onto the original climate forms by the observer (or by the Greenland Ranch superintendent). The purpose of this information is to demonstrate much of the problematic data during the first ten years of record at Greenland Ranch. For some months before about 1916, when the “set max” information was provided, some maximums (on the original forms) were increased to match the set max temperature entered for the previous day. It is assumed that these “write-overs,” or corrections, were made by the USWB. It is possible that these corrections were made by the observer’s supervisor (the Greenland Ranch superintendent), but I do not think that that was the case.
In the tables you will find columns for daily max and min temperature, the daily temperature range, the set max temperature (before 1916), and the precipitation amount. I provide the set max temperature where it is higher than the next day’s maximum temperature, and in the few cases where it is lower than the minimum temperature provided for that day. (I know — that sounds ludicrous, but it did happen). Again, in the temperature tables are the original entries as provided by the observer. The “write-over” entries which match the previous-day set-max temperatures were changed back to the original entry, if that original entry was decipherable.
The average daily maximums and minimums, the monthly means, and the monthly extremes in the tables here may differ some from the values shown on the original forms. They may also differ some from the values to be found in the tables for Greenland Ranch and Death Valley as provided by the Western Region Climate Center (WRCC). Some of the WRCC tables incorporate “corrected” data, and/or temperature data that has been discarded because it appears to be bad data.
These are pdf files. You can make them large enough to read by mousing over the (+) thingie at the bottom and clicking.
GR 1912 daily temps GR 1913 Daily Temps
GR 1916 Daily Temps GR 1917 Daily Temps GR 1918 daily temps
GR 1919 Daily Temps GR 1920 Daily Temps GR 1921 Daily Temps
(MORE UPDATES/January 2021). This entire “part” (Part 2) of the study is quite detailed and quite lengthy. Below, after the overviews and starting with the “Greenland Ranch Data Quality, 1911” section, you will find detailed discussions of the problematic temperature data at Greenland Ranch from 1911 to 1921. This is for those who enjoy getting into the weeds and the particulars and who yearn to really understand why these early temperature records are very suspect. Immediately below, in this updated section, I am going to insert what is hopefully a more succinct and reader-friendly summary of the data issues from 1911 to 1921.
1911 GR Data Overview
Station opened on June 8, 1911
Observer: Thomas Osborn
Set Max data were provided for each observation. The set max data set appear good, but suggest that the time of observation was usually quite a bit later than the official 5 p.m. observations time.
Temperature Data Quality: The GR temperature data for 1911 appear to be very good. The daily ranges are quite large compared to most of the GR temperature record. This strongly suggests that the instruments were above and/or near the “cultivated land” at the ranch, as indicated in the station’s Substation History notation. The evaporative cooling effects of the nearby moist ground surface promotes cooler maximums by a few degrees (F) and also promotes cooler minimums by perhaps 5 to 10 degrees (F) or more most days, compared to nearby locales which are not affected by the cooling effects of the ranch area. The smallest daily range at GR from June to December was 20 degrees! The average daily range was a very healthy 35.7 degrees.
1912 GR Data Overview
Observers: Thomas Osborn for the first 5-6 months. William Weston for about July and August. Oscar Denton for about September to December. Osborn contracted typhoid at the ranch and died in 1912.
Set Max temperatures were provided for each observation. The set max data are not problematic until October 21, when the maximum temperature for the day is lower than the previous-day set max. There were about 12 additional such occurrences from late October through December this year. This problem suggests that the set max temperature entries were “made up,” and/or the maximum thermometer was being reset more than once per day on a regular basis by the new observer, Oscar Denton.
Temperature Data Quality: The GR temperature data for 1912 appear to be very good through September. One exception may be the 120F maximum on May 29, which appears to be an outlier and about 5-10 degrees too high. The minimum temperatures begin to behave in an abnormal fashion by early November. In November and December, the minimums during stretches of many days show little change while slowly decreasing in value. It is hypothesized that the minimum thermometer was not functioning properly and was unusable for the most part, perhaps as early as October 20th. It is hypothesized that the observer may have been resetting the maximum thermometer in the morning to help in an estimation of the minimum temperature. This would explain the beginning of the “set max” temperature discrepancies noted above. The minimum temperatures are not trustworthy from October to December. A minimum of 42F on October 2 is probably a 72F.
All daily temperature ranges were above 20 degrees in 1912, except for one day in July with a range of only 17 degrees. The instrumentation was very likely above and/or near the cultivated ranch land and its significant evaporative cooling effects. The average daily range for 1912 was a very healthy 35.6 degrees.
1913 GR Data Overview
Observer: Oscar Denton
Set-max temperature data are provided for each month, but are “lined-out” on the form for December, 1913. Most months this year have at least a few instances where the set-max temperature is higher than the next-day’s maximum, and/or have “fixed” maximums to match the previous-day set-max temperature.
The overall temperature data quality for 1913 is very poor. The percentage of maximums and minimums which end in a “0” or a “5” is disproportionately high, and is likely indicative of many “estimated” temperature entries. During the spring and summer months are many maximums which appear to be too high as they are not supported by the maximums at surrounding stations (including the July reports of 127F to 134F). The minimum temperatures appear to be especially problematic in the fall and early winter, as they become increasingly and abnormally warm towards December.
Daily ranges of less than 20 degrees become increasingly common after July, and average daily ranges become absurdly low by November and December. The continued large average daily ranges of the first half of 1913 suggest a moist instrument environment, and the very low and very untrustworthy average daily ranges of the last few months of 1913 suggest major problems with the instrumentation and/or observer observing methods.
1914 GR Data Overview
Observer: Oscar Denton
Set max data are provided for each month, but are “lined-out” for January, July-October and December. By fall and early winter, nearly all of the set-max temperatures range from one-to-three degrees lower than the maximum temperature for the day. It appears that all or most set-max temperatures were fabricated for the last few months, at least.
The overall temperature data quality is extremely poor and untrustworthy during 1914. A note for April says that the minimum thermometer had a “split column” and was out of service for three days. The too-warm minimums and too-small daily ranges of late 1913 continued in early 1914 and resurfaced again from October to December, 1914. Most of the maximums of September appear to be estimated, as most are either 109 or 110. Given so many problems, it is difficult to trust ANY of the summer maximums. Some entries may be good, and the extreme monthly maximums are probably good. A high percentage of temperature entries are likely estimates due to the large number of “0” and “5” values.
The very problematic minimum temperature reports contributed to an average daily range of only 23.8 degrees for 1914, or about 12 degrees lower than that for 1911 and 1912. UGH!! Was the thermometer shelter still above or near the cultivated land? The flawed data make it impossible to determine.
1915 GR Data Overview
Observer: Oscar Denton
Set max data are available for January through September only. From June to September, nearly all of the set-max entries are exactly two degrees less than the maximum temperature for day. In other words, the set-max temperature information is of no use, except to provide a well-earned decrease in trust in any and all of the temperature data. On a couple of January dates, the set max temperature was lower than the minimum for the day. There are plenty of write-over maximums during the warm season to match the previous-day set-max temperatures. This is sad, as the set-max data are largely bogus to begin with.
The overall temperature data quality of 1915 is extremely poor. Similar to 1914, the early months of 1915 were afflicted with the absurdly high minimums and small daily ranges. Daily ranges look too small for the bulk of the warm season also, especially if the instrumentation was still above and/or near the cultivated field. The average daily range for 1915 was a paltry 21.6 degrees. Continued minimum thermometer problems likely persisted. Minimums are missing for most of December, and the minimum thermometer was out of service through the following January. The number of “0” and “5” entries soared to 210 in 1915, or about 2 out of every 7 temperatures. (Statistically, this ratio should be close to 2 out of 10.) The “0” and “5” values are believed to represent a good chunk of estimated data, which may or may not be very close to reality.
The warm-season maximums show at least a handful of suspect values. Can most of the monthly maximum temperatures be trusted, and were these from the original shelter location, adjacent to the cultivated field? The best answers that I can provide, given the “look” of the data and how it compares, is “maybe” and “maybe.” Since so much of the data are suspect, it makes it difficult to put much trust into any of the temperature reports.
1916 GR Data Overview
Observer: Oscar Denton
No set-max data provided in 1916.
The overall temperature data quality for 1916 is poor. There were no minimums entered in January due to the out-of-service thermometer, but thereafter the minimums appeared to be much better compared to previous years. Nine consecutive days in June had missing minimums. 241 out of the 690 (or thereabouts) total maximum and minimum temperature entries (35 percent) ended in a “0” or a “5.” There are plenty of timeframes during 1916 (of maybe 2-7 days in length) where the temperatures appear estimated, suspect, or just plain bad. However, the average daily range for the year came up to a reasonable 28.9 degrees. If we can assume that the minimums are accurate for the most part this year, then it may be the case that the instrumentation was now not as close to the cooling effects of the cultivated land at the ranch, and was above the bare ground which was apparent in the station photos circa 1920. I am not ready to hang my hat on that assumption. Like 1915, the monthly maximums in 1916 are probably authentic, but it is very difficult to trust much of the daily maximum temperatures.
1917 GR Data Overview
Observer: Oscar Denton
No set-max data are provided by the observer.
The overall temperature data quality for 1917 at Greenland Ranch is fair, at best. It appears that the minimum thermometer is not problematic, and the absurdly high minimums and small daily ranges of the winters of 1913-14, 1914-15, and 1915-16 are no longer. The temperature data for most of the non-summer months appear reasonable, though with the occasional short spurts of curious data (i.e., the “0” and “5” data and the even-numbers). The quality for the year is deemed as “fair at best” because many of the July and August maximums appear estimated and inauthentic, as they are much too warm compared to surrounding stations. The average daily range of 30.9 degrees in 1917 is about 5 degrees lower than those for 1911 and 1912, suggesting that the instrumentation might be at the photographed but not-officially-documented “bare ground” site by 1917. The current Death Valley station, above bare ground, has an average annual daily range of about 29 degrees. Please see the 1917 section below for more a detailed discussion on the problematic July and August temperatures.
1918 GR Data Overview
Observer: Oscar Denton
No set-max temperatures are provided.
The overall data quality for 1918 at Greenland Ranch is back down to “poor.” Amazingly, 318 out of the 730 maximum and minimum temperatures have a “0” or a “5” as the last digit (44 percent!). It appears that a good number of the temperature entries in 1918 were estimated due to missed observations. No doubt, some estimates were pretty good, and some were pretty bad. The monthly maximum and minimum temperatures are probably good. The average daily range slips down to only 26.6 degrees. If the minimum thermometer was not problematic during 1918 (which seems to be the case), this smallish average daily range suggests that the instrumentation was above bare ground and not cooled significantly by evaporative cooling at the ranch. This researcher also thinks that the decreasing daily range values (compared to 1916 and 1917) are due in part to the increase of estimated entries by the observer, which seem to err on the conservative side. That is, the estimated maximums ran a little lower than reality, and the estimated minimums ran a little higher than reality. A daily minimum of 110F this summer, and one daily July maximum of 85F, are very likely bogus.
1919 GR Data Overview
Observer: Oscar Denton
No set-max temperatures provided.
The overall temperature data quality in 1919 at Greenland Ranch is fair, bordering closely on poor! In 1919, the number of “0” and “5” entries decreased a little, and the average daily range came up to 29.3 degrees. This value is close to what one would expect at a bare-ground station near the ranch. Still, there are plenty of periods with curious temperature patterns: periods with the “0” and “5” and even-number routines during which comparison with surrounding stations tends to become less tight.
1920 GR Data Overview
Observer: Oscar Denton, though the climate forms late in the year show some entries in the handwriting of Victor Ceballos, who replaced Denton in early 1921.
No set-max temperatures are provided.
The overall temperature data quality in 1920 at Greenland Ranch is poor. This year showed an increase of “0” and “5” entries (284 out of 730, or 39 percent), and a decrease in average daily temperature range to only 24.7 degrees. It appears that Denton was estimating a very large number of daily extremes, and that his estimates were conservatively low on the maximums and conservatively high on the minimums. There are too many poor temperature comparisons with surrounding stations this year to provide even a “fair” assessment of the data quality.
1921 GR Data Overview
Observer: Denton for Jan and Feb, Victor Ceballos March to December
No set-max data
I have not looked closely at the 1921 data, but the first two months with Denton as observer continue the trend of numerous estimated temperatures, presumably. The average daily range values increase quite a bit after January to values much greater than during the previous 3 years, which may suggest that Ceballos was a more reliable observer who was missing fewer observations.
Greenland Ranch Data Quality, 1911
Monthly Greenland Ranch and Independence forms for 1911:
Daily maximum temperature charts for Death Valley and select surrounding stations for June to September, 1911:
The Greenland Ranch maximums appear to be non-problematic for the four months covered by the charts above. The daily maximums from Barstow show some curious significant “drops” of daily maximums in July and early August. The max/min reports from Barstow on the 5th, 10th, and 14th were 100/70 in each instance. Since the somewhat low maximums of 100F and these drops (including a drop to 100F on the 17th) are not well-supported by the other stations (and since the Barstow weather was marked as “fair” for each date), it seems likely that the observer missed the actual observation on these dates. The observer likely entered estimated data. The “rounded off to the tens” values of “100” and “70” are obvious clues, or indications, that the temperature entries could be unreliable. Compare how closely the maximums at Barstow mimicked the maximums at Greenland Ranch in June, and then notice how the daily maximums at Barstow bounced around quite a bit in July. A few of the maximums at Barstow in September, 1911, appear to be more than a few degrees too warm, also. The maximums at the remaining stations (Lone Pine, Independence, Columbia and Tonopah) appear to be unproblematic.
All seven “cooperative observers’ meteorological record” forms for Greenland Ranch from June 1911 to December 1911 were signed by “Thos Osborn” (presumably a shortened version of “Thomas Osborn”). The time of observation given is “5 p.m.,” and the “set max” temperature is provided for each day. There are many instances for each month in 1911 where the “set max” is considerably lower than the maximum temperature for the same day. This is not a big deal, but it does suggest that many observations were taken later, or much later, than 5 p.m. More important is that none of the daily maximum temperature values during 1911 are LOWER than the “set max” temperature reading from the previous day —- as standard operating procedure would dictate.
There are a couple of special remarks by the observer during 1911. In June, he notes: “Heavy sand storm from the South on the 29th”; and for September: “Heaviest rain in Death Valley for several years on 29th and 30th.” The months of June, July and August show reports of “0” precipitation, as does December. The “total precipitation” part of the form is blank for October and November. Two days in September show a precipitation amount: “.20” on the 29th and “1.20” on the 30th. According to the form, the rain lasted from 3 p.m. on the 29th to 1 p.m. on the 30th.
The daily maximum and minimum temperature reports for June through December of 1911 appear reliable and unproblematic. Compared to later conditions at the official Death Valley station in Furnace Creek, daily temperature ranges in 1911 were significantly and consistently larger. The difference-maker is, presumably, the ground cover and surrounding ground cover. Today, the station is above bare ground and there is only a small influence from the cooling effects of nearby vegetation and irrigation. In 1911, the instruments were above cultivated land. The irrigated alfalfa crop caused tremendous cooling of the surface air as it passed over Greenland Ranch due to evaporation and transpiration. This likely resulted in a decrease in maximum temperature by a few degrees, especially in summer. And, it almost certainly resulted in a decrease in minimum temperature by 5 to 10 degrees or more. Greatest daily ranges by month at Greenland Ranch, from June to December in 1911, were 43, 44, 47, 45, 44, 48, and 46 degrees. (November 26 had the greatest daily range, with a maximum temperature of 79F and a minimum of temperature of 31F.) The smallest daily temperature ranges by month were 26, 22, 28, 24, 26, 25, and 20 degrees. The average daily ranges for ALL seven months were from 33 to 37 degrees. These are much larger than the average daily range for the year at the current Death Valley station of 28.5 degrees, for the 30-year period from 1981-2010.
Here are the average daily ranges and greatest and smallest daily ranges, and monthly extreme temperatures, at Greenland Ranch (and for Independence, for comparison) for 1911. Also included is the difference in monthly maximum temperatures between Greenland Ranch and Independence, for the high-sun months (June-Sept):
1911 Daily Range Monthly
Ave Gr Sm Mx/Mn
Jun GR 36.2 43 26 120/63 +24
Jun IN 31.5 39 18 96/44
Jul GR 33.5 44 22 122/72 +26
Jul IN 30.0 37 19 96/56
Aug GR 36.7 47 26 122/69 +22
Aug IN 36.5 47 30 100/50
Sep GR 35.2 45 24 118/60 +22
Sep IN 27.2 39 19 96/40
Oct GR 36.3 44 26 100/51
Oct IN 32.3 42 23 81/34
Nov GR 36.9 48 25 90/27
Nov IN 31.5 39 20 72/20
Dec GR 34.6 46 20 80/22
Dec IN 27.8 42 16 63/4
Average daily temperature ranges and monthly greatest daily range data are fairly consistent at desert stations, from year to year and from decade to decade, when there is no change in the environment around the temperature equipment. Daily temperature range data can be very useful for detecting suspect temperature data. Sudden, persistent, and/or significant changes in daily range values often indicate problematic reports (if they cannot otherwise be explained by the weather, and if the change is not associated with movement of the instruments or a change in the ground cover or local ventilation, etc.).
Note that there were no days from June to December in 1911 with daily ranges of less than 20 degrees at Greenland Ranch. Rare are the days in Death Valley when cloud cover allows little to no sunshine, and rare are the nights when cloud cover is thick enough to prevent plenty of surface cooling due to radiational heat loss. Dry air at low levels in Death Valley, plenty of clear to mostly clear skies, and the station’s location near the bottom of the basin contribute to and promote the consistently moderate-to-large daily ranges. The narrowest daily ranges would be during rainy weather, overcast skies, and/or windy conditions. If the wind blows all night and prevents a surface radiation inversion from developing, then the overnight temperature will remain comparatively very warm. In these circumstances, the maximum temperature for the following afternoon is typically no more than 15 to 20 degrees warmer than the minimum. Typically, though, there are lulls in the wind through Furnace Creek for parts of the night which allow a shallow inversion to develop. The result is a temperature at shelter level that is 5, 10, even sometimes 20 degrees cooler than the “free air” temperature one or two hundred feet above.
A locally moist ground surface in a desert environment promotes even lower minimums (and larger daily temperature ranges) through evaporative cooling. The Greenland Ranch instrument shelter was initially installed above an alfalfa sod. Plenty of irrigation was necessary to grow the sod. This is a major difference compared to the bare (and unirrigated!) ground below and around the shelter today. From photos, the floor of the Greenland Ranch instrument shelter looked to be a foot or so closer to the ground than the current “Death Valley” shelter at Furnace Creek, and this would also promote somewhat lower minimums comparably.
So, again, the placement of the Greenland Ranch instrument shelter above the alfalfa sod resulted in significantly larger daily temperature ranges. Though maximum temperatures were likely tempered (i.e. lower) by perhaps two to three degrees due to the placement over sod (compared to an alternate placement above bare ground and away from any evaporative cooling effects of irrigation and vegetation); the minimum temperatures were substantially cooler comparably.
Today, daily ranges of 40 degrees or more are somewhat rare at Death Valley. Daily ranges of 40 degrees or more were not uncommon back in 1911 and 1912, when the Greenland Ranch shelter was above sod.
Greenland Ranch Data Quality, 1912
(revised and updated in January, 2021)
Monthly Greenland Ranch forms for 1912:
Monthly Independence forms 1912:
Daily maximum temperature charts for Death Valley and select surrounding stations for May to September, 1912:
The charts show what are likely some problematic, or suspect maximums during these summer months. In May, 1912, note how the Lone Pine maximums rise in a very gradual and gentle manner. Significant regional cool-downs on several occasions, especially on May 20 and May 25, are not apparent in the Lone Pine record. It is likely that the observer was not around to take the daily observations during parts of this month, and entered “ballpark” values which seemed reasonable and “fit” the general trend for the month. Similar “smoothing” of the Lone Pine maximums from day-to-day is also evident in the charts for June, July and September. Thus, for Lone Pine this summer, the monthly maximum temperatures are likely trustworthy, but the daily maximums simply cannot be trusted.
There was a warm spell for the region around May 29, but the Greenland Ranch maximum jumped up quite a bit more than the maximums at the other stations. The maximums at the surrounding stations on May 29 strongly suggest that the 120F maximum at Greenland Ranch for that date is at least 6 to 8 degrees too warm. What caused that? The original form for May 1912 appears to be in the writing of Thomas Osborn, the original Greenland Ranch observer. The form for May 1912 was the last one that Osborn filled out, though. Osborn died, likely from heat-related health problems, soon afterwards. It is possible that Denton was already serving as a replacement caretaker and observer as early as late May, 1912. It is possible that Denton artificially inflated the maximum temperature at Greenland Ranch on May 29, 1912. Will we ever know for certain? I doubt it!
The Greenland Ranch observer forms were signed “Thos Osborn” for the first five months of 1912. The form for June is signed “Thos Osborn,” but his signature and the data entries were by someone else, based on the change in handwriting. That “someone else” may have been superintendent Fred Corkhill, as the initials “FWC” are below the signature on the June form. (It may be that Corkhill filled out a more legible replacement form to mail in to the Weather Bureau.) The data reports for July and August appear to be in somewhat different handwriting than that by Thomas Osborn, and the forms are signed “Wm Meston for Thos. Osborn.” The records for September and the first 17 days of October again appear to have somewhat different handwriting than that of Thomas Osborn, and the forms for both months, and for November and December of 1912, are signed “OA Denton.” The entries in Denton’s handwriting begin on October 18, 1912. There were no special remarks entered on the forms for 1912 (though, as for 1911, the “character of the day” entries show clear, cloudy, wind, shower, etc.).
Seven months of 1912 had “0” or “0.00” precipitation totals at Greenland Ranch, according to the observers’ record. Two days each in April and May had a shower, but total precipitation for those months was only a trace. March had two days with measurable precipitation: “.7” on the 9th and “.4” on the 29th, as written on the forms, for a monthly total of 1.10 inches. There were two more days with measurable rain at Greenland Ranch while Thomas Osborn was observer (apparently through October 17, 1912). The entry for July 30 is “1/10”, and for October 5th is “.20”. Oscar Denton recorded zero precipitation through the remainder of 1912.
Thomas Osborn entered measurable precipitation amounts on six days during his (assumed) tenure as observer from June 8, 1911 through October 17, 1912 (though he may have only been observer through about May of 1912). All or most of these were to the tenth of an inch. Of course, the standard rain gage at the station would permit measurements to the nearest hundredth of an inch. Further, the precipitation entries on the Greenland Ranch forms during this time frame appear NOT to be in the handwriting style of the person (presumably, Thomas Osborn) who entered the daily temperature extremes and the “character of the day.”
The daily maximum and minimum temperature reports for 1912 appear reasonable through September. One exception may the May 29 maximum of 120F, which appears to be too high based on surrounding maximums. Denton may have already been filling in as the weather observer on occasion by late May, 1912. Did he artificially enhance the maximum on May 29, 1912? Curiously, three consecutive maximums (from May 28 to May 30) end in a zero (110, 120 and 110). It later became commonplace to find an unusual number of these “zero” entries during Denton’s tenure, when it appeared that he was “filling in the blanks” for periods with missed observations.
During November and December, the Greenland Ranch minimums behave very strangely. There is a relatively smooth and consistent decrease in minimum temperature these months. The change, or behavior, of the minimums does not appear natural. In addition, there is the occasional maximum temperature which is lower than the previous-day set-max temperature after Denton took over.
The following table provides the average daily range, the greatest and smallest daily ranges, and the temperature extremes for each month in 1912, at both Greenland Ranch and at Independence. Also provided is the difference in monthly maximum temperature between the two stations for the high-sun months from April to September:
1912 Daily Range Monthly
Ave Gr Sm Mx/Mn
Jan GR 39.5 50 29 85/22
Jan IN 31.4 42 23 68/12
Feb GR 41.3 53 20 88/33
Feb IN 32.1 41 17 73/18
Mar GR 32.0 44 21 91/38
Mar IN 24.6 34 8 64/22
Apr GR 34.2 47 23 98/36 +24
Apr IN 26.5 39 14 74/25
May GR 35.5 51 26 120/55 +28
May IN 28.9 37 13 92/37
Jun GR 38.5 49 24 120/57 +23
Jun IN 31.6 39 16 97/32
Jul GR 31.1 48 17 118/67 +20
Jul IN 30.4 39 19 98/52
Aug GR 30.7 43 23 120/68 +20
Aug IN 33.0 39 24 100/51
Sep GR 35.4 48 21 112/56 +21
Sep IN 30.5 38 20 91/39
Oct GR 33.0 50 21 99/42
Oct IN 29.4 40 14 80/25
Nov GR 39.6 51 22 87/32
Nov IN 30.1 42 16 75/22
Dec GR 37.7 51 27 82/21
Dec IN 29.0 45 15 68/12
(Keep in mind that observation times in the late afternoon are typically just an hour or two or three after the high temperature of the day is reached, when the temperature can still be somewhat close to the maximum. This works “against” recording small daily ranges. For instance, if the maximum for the day is 80F, the “set max” at 5 p.m. is 75F, the overnight minimum is 50F, and the following daytime maximum is only 55F due to clouds and rain; the daily range would still be 25 degrees, and not 5 degrees, because the 24-hour maximum was 75F. Had the observation time been at midnight, the “smallest daily range” figures would no doubt be lower for most months, and the average daily range would be a little lower.)
The greatest daily range for the year was 53 degrees, on February 7, with maximum and minimum temperatures of 88 and 35F. All of the days of the year had temperature ranges of 20 degrees or larger, except for July 17th, which had a daily range of only 17 degrees (109/92F). The paucity of daily ranges of 20 or less during 1911 and 1912 is rather amazing.
Like 1911, the “set max” temperatures during the first few months of 1912 seemed to run cooler than expected given the 5 p.m. observation time. This suggests that Osborn often read and reset the instruments later than 5 p.m. During the summer months, the differences between the day’s maximum and “set max” temperatures decreased. This trend of smaller differences continued into November and December, when Denton was the observer.
The minimum of 42F on October 2 at Greenland Ranch is suspect (too low). Given the cloudy and rainy conditions in the area around this date, the October 2 minimum may be 72F. The entry is not clear on the original form, but may well be 72F.
Through October 17, 1912, there were no more than a couple of instances in which the day’s maximum temperature was lower than the “set max” temperature on the previous day at Greenland Ranch. The hand-written entries show that Oscar Denton replaced Thomas Osborn as station observer on October 18; and on October 21, the maximum temperature of 89F was lower than the “set max” temperature of 92F on October 20. On the 27th of October, the maximum of 90F was two degrees lower than the “set max” temperature on the previous day, and on the 29th of October the maximum of 74F was four degrees lower than the “set max” temperature on the previous day. Obviously, if the maximum thermometer reads a certain temperature just after it is reset, it will not indicate a lower value the next time it is read (unless it has been reset again in the meantime). These instances (i.e. “errors”) are not of significant consequence with regard to the long-term averages at Greenland Ranch, they would not have any effect with regard to extreme maximum temperatures or monthly maximums in Greenland Ranch’s record, and it could be argued that the “adjusted” maximums more accurately reflect conditions at Furnace Creek and are welcome. However, the improper observing practice does say something about the station observer. Right off the bat, the reports provided by Denton are problematic from an observational standpoint. The reports by this observer require greater scrutiny compared to those from an experienced, well-trained and conscientious observer.
In November of 1912, there were four instances in which the day’s maximum temperature report was lower than the previous day’s “set max” temperature, and in December this occurred on six days. One can only speculate as to the reason for these apparent “errors”. Was Denton routinely resetting the maximum thermometer during the morning as well as at the 5 p.m. observation time? That seems to be the case. Denton may have been especially interested in the daily temperatures at the ranch, and he may have wanted to eliminate the chance of a “previous-day” maximum temperature due to a high “set max” reading on the maximum thermometer. By resetting the maximum thermometer in the morning, Denton would be able to record the “real” high temperature for the day. It would eliminate the chance that the “set max” temperature from the previous afternoon would wind up as the maximum for the day. Again, the consequences of this particular practice are largely inconsequential. Five p.m. is relatively close to the warmest time of the day, and is not a very good time for the observation if one wishes to minimize the chance for a “previous-day” maximum. Denton apparently got around this occasional nuisance by resetting the maximum thermometer more than once per day. But, he still should have entered the “set max” temperature from the previous day as the current day’s maximum if the maximum temperature reading at 5 p.m. was lower!
Updated thoughts — January 2021!
My hypothesis that Denton may have been resetting the maximum thermometer more than once per day because he wanted to avoid the possibility of a dreaded (sarcasm) “previous-day”maximum is probably not correct. A close examination of the minimums at Greenland Ranch during November and December shows very peculiar behavior, as the values show a relatively slow and steady and unnatural decline. Given the sudden number of maximums which were lower than the previous-day set-max temperatures about this same time, it may be that there was a problem with the minimum thermometer already. It may be that the minimum thermometer was broken, and it may be that the maximum thermometer was reset in the morning in order for the observer to know the current temperature. If the maximum thermometer was getting reset twice per day, then that would explain the “set max” temperature discrepancies. If Denton was resetting the maximum thermometer in the morning in order to obtain the minimum for the day, it would appear (in November and December of 1912, at least) that he just used it as a guide, as the peculiar nature of the minimums strongly suggest that most or all are estimates. Unfortunately, the minimums of Nov and Dec in 1912 at Greenland Ranch are not trustworthy. The average and even the extreme minimum data for these months should be discarded and not used to determine any long-term climate records.
Greenland Ranch Data Quality, 1913
Monthly Greenland Ranch forms for 1913:
Monthly forms for Independence, CA — 1913
Daily maximum temperature charts for Death Valley and select surrounding stations for April to December, 1913:
Note again that the chart for April 1913 (above) shows the same issue at Lone Pine which was discussed above for 1912. The daily maximums seem to be “smoothed over,” notably a couple of cool periods early in the month. The same is true during May at Lone Pine, but the remaining summer months appear more reliable (or should I say “less unreliable?” The bottom line here is that it is NOT advisable to directly compare daily Greenland Ranch maximums with the daily maximums at Lone Pine, at least without a large amount of care and caution! This is unfortunate, because Lone Pine was the closest operating weather station to Greenland Ranch during the summer of 1913, when the record 134F maximum was recorded in Death Valley! Fortunately, the station at Independence, with its reliable USWB observers, was only about 10 miles farther from Greenland Ranch compared to Lone Pine. Independence was the closest reliable, long-term weather station to Greenland Ranch in 1913.
When Oscar Denton became the ranch caretaker and observer at Greenland Ranch in late 1912, the reliability of the maximum temperature record here became increasingly problematic. A quick look through the charts above show plenty of suspect maximums at Greenland Ranch during 1913, including:
–dubious (too warm) maximums of 100F on April 2 and April 8, and maximums which appear too warm by about five degrees (F) during the final week of April
–dubious (too warm) maximums of 110F on May 4th and 120F on May 25th (both at least 5 degrees F too warm)
–dubious (too warm) maximums during the last week of June, especially the 118F on June 26, which is at least 12 degrees too warm.
–too warm maximums by 5-10 degrees F during the first half of July
–too warm maximums by about 3-5 F degrees at Greenland Ranch for much of August, 1913.
–for September, 1913, a very dubious (too warm) maximum of 118F on September 3rd (23 degrees higher than the maximums at Barstow and Las Vegas!) Note — the entry looks like “118” but may be “108.” The maximum shown for the entire month is 116F. This comparison for September 3 should probably be disregarded.
Recall (from Part 1) that maximum temperatures at Greenland Ranch during the spring of 1913 appeared to creep upwards when compared to surrounding stations. It is not apparent that there was a certain day that we can point to and say “here—here is where things seemed to go off the tracks for good.” The upward trend in maximums was a bit erratic. With this in mind, let’s dissect the record at Greenland Ranch during 1913.
The twelve “observer’ meteorological record” forms for 1913 were signed “OA Denton”. The January form provided a remark (“Unusually cold for Death Valley”) but the others had no entries in the “remarks” section.
Continuing the trend of 1911 and 1912, the precipitation reports were peculiar at Greenland Ranch in 1913. There were 16 days with measurable precipitation amounts. (There were no reports of “trace”.) Of these, 12 were to the tenth of an inch, and the remaining four were all “.o1”. Six of the precipitation entries were written as fractions (e.g. 1/10, 2/10, and 3/10 for July 20, 21 and 22). And, it again appears that at least some of the precipitation amounts were entered onto the forms by someone other then the person who entered the temperature data. For the first 31 months of record at Greenland Ranch, all of the days with measurable precipitation entries (22 of them) had amounts of either one-hundredth of an inch, or an amount rounded off to the nearest tenth of an inch.
The Greenland Ranch temperature record is tremendously flawed for parts of 1913, and the July maximums are not even the most egregious of the problems. The minor observer error issues with regard to the “set max” and next-day maximum temperature persisted. Month by month during 1913:
Jan: three maximum temperatures appear to be “fixed” to match the previous day’s “set max”. (Someone wrote over the original entry.) The corrections were likely made by the Weather Bureau after the form was mailed from Ryan. We’ll call these entries “fixed max temps”, “FMTs” for brevity.
Feb: one instance of the day’s maximum lower than the previous day’s 5 p.m. “set max” temperature. We’ll call these “bad max temps”, or “BMTs” for brevity. Also, one day “fixed”.
Mar: three BMTs The total difference of 33 degrees resulted in an average maximum for the month that was a full degree lower than would have been realized had proper observing practices been followed.
Apr: no BMTs, but 5 or 6 FMTs. It appears that five or six BMTs were “fixed” by someone who reviewed the submitted form.
May: 6 FMTs
June: 2 BMTs
July: There were just a couple of “set max” issues this month, the month with the 130F reports. On the 13th, a maximum of 131F was recorded, and a 5 p.m. “set max” temperature of 127F was entered. The maximum for the next day, the 14th, is 127F, but it appears that the “7” is covering another number. On the 18th, the maximum of 108F is seven degrees lower than the “set max” for the 17th. On the 20th, it appears again that the entry of “105” was written on top of an earlier entry, in order to match the “set max” from the day prior. The same goes for the 26th, when 104F looks to be covering a “101”.
Aug: Some of entries on the early record forms from Greenland Ranch are difficult to decipher. In some cases the penmanship is poor, “9’s look like “4’s” (and vice versa), parts of numbers stray into adjacent entries, etc. But, most of all, the copies are often very poor, with parts of numbers just fading away. (Occasionally, the “set max” information is very helpful to determine what the maximum temperature entry is.) The NCDC original copy for Greenland Ranch for August, 1913, is not strong, and all of the temperature entries are “write-overs” in order to make the entries easily readable. On the 9th, however, the “set max” value is “48”. This is for a summer day with a maximum temperature of 109F, so someone goofed here. On the 27th, the maximum of 110F is nine degrees lower than the “set max” from the 26th, and seven degrees lower than the “set max” of 117F on the 27th—another goof—but by whom? The dark write-overs by, presumably, a Weather Bureau employee, totally masks the original entries. Maybe the maximum on the 27th was actually 120F.
Sep: 3 BMTs
Oct: 5 BMTs
Nov: 6 BMTs
Dec: 3 BMTs…and there is a line through all of the “set max” temperature entries. It is presumed that a Weather Bureau reviewer determined that the “set max” data should be disregarded. As we’ll see shortly, much of the December 1913 temperature data from Greenland Ranch are problematic.
Again, a “BMT” is a (bad) maximum temperature entry that is lower than the previous day’s 5 p.m. “set max” temperature. This is an observer error, and these were rather common during 1913. Some of the forms for 1913 show “write-overs” in the maximum column to match the previous-day “set max” entries. I suspect that the corrections were made by the Weather Bureau after the forms were mailed from the post office at Ryan, CA. If that is the case, then both the observer (Denton) and the superintendent (Corkhill) are responsible for the errors, and the problem suggests that the observer (or observers) was improperly trained, unconcerned about proper procedure, or perhaps TOO enthusiastic as observer. My hunch is that Denton paid plenty of attention to the thermometers in the shelter, he disliked previous-day maximums, and he regularly reset the maximum thermometer in the mornings so that he would know the “real” high temperature for the day.
The greatest daily range values for 1913 at Greenland Ranch not only suggest, but they virtually scream out loudly, that the temperature record for much of the year CAN NOT be trusted! Recall that for 1911 and 1912, average daily ranges for all of the months were roughly from 31 to 41 degrees. These figures are quite reasonable for a station near the bottom of a deep desert basin. Greatest daily ranges were commonly near 50 degrees, and the smallest daily ranges for each month were generally 20 to 30 degrees. Only one day from June 1911 through December 1912 had a daily range of less than 20 degrees. By the end of 1913, daily ranges of 20 degrees or less were the norm, according to the record! Something changed drastically at Greenland Ranch from 1912 to 1913, besides the observer!
In June, 1915, G.H. Willson provided a write-up in Monthly Weather Review on the establishment of the Greenland Ranch weather station and the record maximum temperatures in July, 1913. He also provided a table giving the average maximum and minimum and mean monthly temperatures for the first four years of record:
A quick perusal of the “minimum temperature” and “mean minimum temperature” parts show that minimums, particularly in the cooler months, increased tremendously (and unrealistically) during the second half of this four-year period as compared to the first half. The reported average minimums at Greenland Ranch appear to be FAR too warm during the winters of 1913-14 and 1914-15. Why did Willson fail to note the inconsistent minimum temperature record at Greenland Ranch?
Of course, the steep upward trend in minimums during the last few months of 1913 at Greenland Ranch had a profound effect on the daily range values. The following table provides the average daily range, the greatest and smallest daily ranges, and the temperature extremes for each month in 1913, at both Greenland Ranch and at Independence. Also provided is the difference in monthly maximum temperature between the two stations for the high-sun months from April to September:
1913 Daily Range Monthly
Ave Gr Sm Mx/Mn
Jan GR 38.6 48 21 82/15
Jan IN 25.1 35 3 65/4
Feb GR 30.8 47 16 90/30
Feb IN 22.6 35 7 67/16
Mar GR 36.5 49 26 98/36
Mar IN 28.3 37 12 78/14
Apr GR 38.0 52 27 109/47 +25
Apr IN 29.1 38 6 84/27
May GR 37.2 52 27 120/52 +31
May IN 30.8 37 11 89/29
Jun GR 35.1 46 24 119/60 +27
Jun IN 30.0 39 12 92/42
Jul GR 35.6 53 14 134/70 +31
Jul IN 28.2 41 16 103/51
Aug GR 32.1 46 20 124/74 +28
Aug IN 28.8 38 14 96/50
Sep GR 29.7 39 13 116/66 +22
Sep IN 30.4 38 19 94/39
Oct GR 24.4 33 14 105/60
Oct IN 30.3 38 20 82/29
Nov GR 12.4 22 1 90/58
Nov IN 23.9 36 6 76/24
Dec GR 7.9 16 3 74/52
Dec IN 21.6 41 8 61/18
Average daily ranges began to decline during the summer, and by October 1913 the daily ranges appeared truly suspect. Octobers are typically dry, mostly clear, and not as breezy as during the summer. This promotes fairly sizable daily ranges at desert stations near the bottom of deep basins, such as at Furnace Creek. The first 20 Octobers at Greenland Ranch had an average daily range of 32.6 degrees (90.3/57.7F). The average daily range in October 1913 was only 24.4 degrees, and the greatest daily range was only 33 degrees. This might be explained away if the month were unusually cloudy and wet, but it was not. There was no rain reported at Greenland Ranch this month, and 24 days were characterized as “clear” by the observer. The average daily ranges in the previous two Octobers were 36 and 33 degrees…why was it only 24 degrees in October 1913? The answer is easy: minimum temperatures, or should I say, “the minimum temperatures which were recorded,” were exceptionally warm! The mean minimum of 70.2F was about 13 degrees warmer than average. And why was the average minimum so warm? The average maximum was only four degrees above the (1911 to 1930) average. At Independence, the average minimum of 42.3F was one degree below its average through 1930. At Barstow, the average minimum for October 1913 was 58.9F, 11.6 degrees above average. So, given Barstow’s very warm nights, perhaps October 1913 was unusually humid and/or breezy in Death Valley, and maybe it is unfair or unwarranted to summarily dismiss the warm minimums —- if Barstow’s record is legitimate.
The average maximum and minimum temperatures at Greenland Ranch in November 1913 were 78.6 and 66.2F, for an average daily range of only 12.4 degrees. The previous two Novembers had average daily ranges of 37 and 40 degrees! Certainly, there is a problem with the Greenland Ranch data now, and it appears that the minimum temperatures are much too warm. The average minimum of 66.2F in November 1913 is 20 degrees above average. At Independence, the average minimum this month was 0.3 degrees above average, and at Barstow the average minimum was 16.9 degrees above average! This is a bit strange. The month was cloudy and wet. Is it possible that the very small daily ranges at Greenland Ranch in November 1913 are legitimate? November 11 was clear with extremes of 89 and 80F. This is the highest minimum for any November day at Greenland Ranch (through 1960), and the next warmest November minimum is 75F (on November 8, 1913). For the period from 1911 to 1960, 13 of the 30 dates for warmest minimum belong to 1913, and seven belong to 1914. This is according to the daily summary stats on the Western Region Climate Center web site. Were the nights exceptionally warm for the month, or are the Greenland Ranch minimums for November 1913 garbage?
The ridiculously miniscule daily ranges at Greenland Ranch for December, 1913, and continuing into the following winter months, permit an easy answer. The minimum temperature reports from Greenland Ranch begin to appear too warm in September or October, and by December they are, without a doubt, much too warm —- for reasons that are difficult or impossible to determine. The average maximum and minimum temperatures at Greenland Ranch in December, 1913, were 65.7F and 57.8F, giving an average daily range of 7.9 degrees. The average daily range for December from 1911 to 1930 was 27.6 degrees (65.8/38.2). (If the suspect data from December 1913 and December 1914 are disregarded, this value edges up to 29.5 degrees.) As was the case in November, the average minimum in December 1913 was about 20 degrees above normal! The averages and departures this month:
Greenland Ranch 65.7/57.8 (-0.1/+19.6) Greatest range 16
Independence 46.4/24.8 (-5.7/-2.1) Greatest range 41
Barstow 64.6/40.1 (+4.6/+10.0) Greatest range 39
Las Vegas 57.6/31.2 (-2.4/-0.4) Greatest range 37
Needles 65.5/44.5 (+2.6/+4.7) Greatest range 28
Tonopah 34.0/23.8 (-6.0/-1.4) Greatest range 21
Average daily ranges at stations surrounding Greenland Ranch were near-to-a-little smaller than average this month, about 21 to 26 degrees at the basin stations. Greenland Ranch’s average range of 8 degrees is even smaller than at Tonopah, an extremely well-exposed ridge-top town at 6000 feet elevation! (Note: the relatively warm minimums at Barstow from October to December of 1913 may be problematic.)
During December, 1913, maximums at Greenland Ranch ranged from 60F to 74F, and minimums ranged from 52F to 65F. The maximums at Greenland Ranch do not appear unreasonable at a glance, but there were a couple of cool spells during which one might expect cooler maximums at Greenland Ranch. For instance, Las Vegas showed a high of 42F on Christmas, while Greenland Ranch reached 60F. Tonopah had highs of only 22F on the 19th and 20th, and Greenland Ranch managed 63F and 69F.
A single line is drawn through the column with the “set max” data on the Greenland Ranch form for December 1913. Perhaps someone looked at the data for December 1 — a maximum of 68F, a minimum of 65F, and a “set max” of 64F! Ooops! The “set max” temperatures are unusually consistent with regard to the difference from the maximum this month. Whereas in the previous Decembers the “set max” temperatures varied quite a bit from the maximum (routinely lower by 5 to 15 degrees or more), in December 1913 the “set max” was 1 to 4 degrees lower than the day’s maximum on all but a few occasions.
The daily range at Greenland Ranch was only three degrees on both December 1 and 2 of 1913. The smallest daily ranges for the previous two Decembers were 20 and 27 degrees. There were 14 days during December 1913 with daily ranges of six degrees or smaller!
The high minimums at Greenland Ranch in December 1913 averaged 28 degrees warmer than the minimums there in December 1912! (At Independence, one of the closest long-term stations to Greenland Ranch at the time, the average minimum temperature for December 1912 was only two degrees warmer than for December 1913.) The low for the month at Greenland Ranch was a balmy 52F (and 57F was the minimum on 12 days!). There was a repeat of very warm minimums twelve months later, and Greenland Ranch again had a low of 52F in December, 1914. The next highest monthly minimum for the 50 Greenland Ranch Decembers from 1911 through 1960 was 35F!
There are numerous ways to show that the temperature data from Greenland Ranch in December, 1913, are deeply flawed. The minimums, at least, are grossly higher than those which would be expected, and this continues a trend which began in October, if not a little earlier. There is zero chance that the air in the Greenland Ranch thermometer shelter varied in temperature by an average of only 8 degrees during these 31 days. Why do the reports show this, though? How did this occur? Was the minimum thermometer malfunctioning? If so, why didn’t the observer, Oscar Denton, notice it? Why didn’t he remedy it, why didn’t he report it? Why didn’t the superintendent of the ranch, Fred Corkhill, see any problem and fix it?
There are numerous “things” that can happen at a weather station which could, or would, result in problematic temperature data. The liquid in the thermometers can separate, the indicator in the minimum thermometer can get stuck or lodged in the wrong part of the inner glass tube, alcohol in the minimum thermometer can slowly collect in the reservoir at the top end of the tube, causing incorrect readings…among other scenarios. There is also the human element to consider. Is the observer reliable and competent? Is the observer one who understands and welcomes the task, and who strives to produce the most accurate and reliable climate record possible? Or, does the observer find the “job” as just another thing he has to do that he would prefer not to have to do? On the other extreme, is the observer one who is obsessed with the activity? Does he “bend the rules” or stray from proper procedure to some degree to “make things right”, in his mind?
It is difficult to conceive that Denton did not know that the minimum temperature reports (at least) were compromised. Whether there was an instrument problem or not, how can the observer routinely record obvious bogus minimums in December in the 50s and 60s, especially when he was recording 20s and 30s during the previous December? How does that happen? Was there a massive separation in the minimum thermometer liquid and did Denton really think that he was reading a properly functioning instrument? Did the indicator in the minimum thermometer wind up on the wrong side of the alcohol in the thermometer somehow? Was the minimum thermometer left at a steep angle on a consistent basis which allowed the indicator to creep down, resulting in warmer readings?
If there was a problem with the instruments and the observer did not notice or understand, and/or did not care, and he still recorded the problematic data, then that casts serious doubt upon the adequacy of the observer’s credibility, capability and/or conscientiousness. If there was in fact NO problem with the instruments or the thermometers in the Greenland Ranch shelter during December, 1913, then how did these warm minimums come to be? The maximums and minimums at Greenland Ranch in December, 1913, look like readings which are from the inside of a house or building! Did Denton move the shelter, and/or the instruments, inside? How else might miniscule daily temperature ranges such as these come about in Death Valley?
On December 27, the maximum temperature was 66F and the “set max” temperature (at or around 5 p.m.) was 62F. The minimum for the next day, in the morning (presumably) was 54F, and then a maximum of 60F on the 28th was recorded. (The “set max” on the 28th was 59F.) Here is a case where the maximum is lower than the “set max” from the previous day, which suggests that the observer was resetting the maximum thermometer more than once a day. This occurred a couple of other times in December 1913. If the observer is interested enough in the high temperature from day to day and he feels compelled to do this, then it is probably a stretch to suggest that the observer had the thermometers inside a building (which could explain the persistently small daily ranges). But, perhaps the maximum thermometer was in its proper place and functioning normally inside of the shelter, while the minimum thermometer was somewhere else, in a warmer and totally unrepresentative spot. Or, perhaps the minimum thermometer was hopelessly broken and of no use at all. If this is the case, then maybe Denton was looking at another thermometer at the ranch to estimate minimums, or perhaps he was spinning the maximum thermometer late in the morning, or at lunch time, and using THAT reading as the minimum for the day! It really sounds ludicrous, but the fact is that Denton entered obviously untrustworthy minimum temperatures for months at a time, day after day. He HAD to know that the data that he was recording was not realized according to standardized procedures. Either that, or Denton really did lack an understanding of what was required to procure reliable and trustworthy temperature data. Recall that his boss, ranch superintendent Fred Corkhill, wrote in a letter to the Weather Bureau that Denton was unable to figure the humidity value “from the tables” after taking the wet and dry bulb readings. Corkhill said that checking the hygrograph against the sling psychrometer was “a little out of his line.” If this is the case, then Denton was probably not fit to be the station observer. The daily weather observation at Greenland Ranch likely ranked near the bottom on the list of daily duties at the ranch. Corkhill probably did not have much of a choice when it came to selecting the weather observer. Since Corkhill apparently was not at the ranch regularly enough to take the observations, and since Denton was willing to sweat out the entire summer at the ranch, Denton was the guy to replace Osborn.
Oscar Denton signed the observer forms from Greenland Ranch beginning in September, 1912. It appears, based on his handwriting, that the daily entries by him began on October 18, 1912. Denton’s first full year as observer was 1913, and a close look at the record from 1913 uncovers a handful of errors, or “issues”, large and small. Among these are the “set-max” issues, the curious precipitation amounts and entries, some seemingly inflated warm-season maximums (including those during the record week in July), and the problematic warm minimums from October through December. It is difficult to determine just how and why the Greenland Ranch temperature record is so poor during 1913. The answer may be as simple as “the observer ignored proper procedure and many (extremely suspect) temperature observations were simply contrived.”
Greenland Ranch Data Quality, 1914
Monthly Greenland Ranch forms for 1914:
Monthly Forms for Independence, 1914
Daily maximum temperature charts for Death Valley and select surrounding stations for all months, 1914:
A perusal of the monthly charts for the Death Valley-area in 1914 reveal the following likely problematic maximums:
— highs of 86F and 89F on Jan 2 and Jan 4, respectively. Comparing maximums among regional stations in the winter months is very hazardous due to cold air which commonly becomes trapped in the desert basins. Some low-elevation basin stations can have cooler maximums than stations that are thousands of feet higher. Some high-elevation stations might have snow cover which could promote lower-than-expected maximums. Even so, these upper-80s maxima at Greenland Ranch are difficult to explain and are quite anomalous for early January. It is possible that someone “wrote over” the original entries with incorrect values on these two dates.
–a big jump in temperature from March 14 to 15 (from 82F to 97F) is not supported by the other stations (which saw increases of about 2 to 6 degrees). It may be because the maximum on the 14th at Greenland Ranch is incorrect (too low). The Greenland Ranch maximums in March are supported by the maximums at Independence.
–Lone Pine maximums again appear abnormally “smoothed out” for the last 10 days in May 1914. Greenland Ranch maximums look okay in April and May.
–on June 19, the maximum at Greenland Ranch jumps up 7 degrees (to 112F) from the previous day while all other stations show a decrease in the maximum. The other June maximums appear okay at GR.
–the maximums at GR in July appear to run a few degrees warmer than expected (at least compared to differences between other stations for the Julys of 1911 and 1912). The maximums at Columbia from July 14 to 24 appear to be too warm by 6-9 degrees. Perhaps a new or substitute observer was not reading the thermometer properly at Columbia, similar to the problem leading to the world record-high in Libya (of 136F) which was thrown out.
–GR maximums appear to be a few degrees too warm from about August 5 to 17
–GR maximums are highly suspect beginning September 12 through the end of the month. It appears that the observer was absent during this period and estimated three weeks of maximums.
–several maximums at GR during October appear to be too cool. Again, the observer may have been absent for a few periods, and estimated maximums for a handful of days when he returned.
–November maximums at GR look okay, but its December maximums seem much too warm for most of the month, especially after the 9th.
Unreliable and unrealistically warm minimums continued into early 1914. Minimums appeared relatively reasonable during the warm months; and, as occurred in 1913, were much warmer than expected from October through December. In April, 1914, a remark on the observers’ record form states: “Minimum thermometer out of service 3 days on account of a split column.”
The following table provides the average daily range, the greatest and smallest daily ranges, and the temperature extremes for each month in 1914, at both Greenland Ranch and at Independence. Also provided is the difference in monthly maximum temperature between the two stations for the high-sun months from April to September:
1914 Daily Range Monthly
Ave Gr Sm Mx/Mn
Jan GR 7.3 17 2 73/52
Jan IN 19.0 29 4 57/17
Feb GR 15.6 26 4 85/47
Feb IN 26.0 36 10 64/20
Mar GR 25.7 32 13 98/53
Mar IN 30.3 37 18 78/26
Apr GR 26.2 42 14 104/50 +20
Apr IN 28.0 40 16 84/29
May GR 28.9 45 22 116/53 +26
May IN 29.8 37 17 90/35
Jun GR 30.9 44 21 124/60 +23
Jun IN 30.7 39 10 101/39
Jul GR 30.1 40 14 122/76 +26
Jul IN 31.3 39 26 96/53
Aug GR 31.7 41 21 126/70 +26
Aug IN 33.3 39 25 100/50
Sep GR 28.1 50 23 112/60 +24
Sep IN 32.1 40 18 88/39
Oct GR 19.2 28 11 101/68
Oct IN 30.3 42 19 82/38
Nov GR 17.8 24 10 91/57
Nov IN 33.5 40 25 74/14
Dec GR 13.5 23 8 82/52
Dec IN 23.1 30 10 50/3
Recall that the average daily temperature ranges during the winters of 1911-12 and 1912-13 were approximately 35 to 40 degrees. Average daily ranges plummeted to 7.9 degrees in December, 1913, and to 7.3 degrees in January, 1914. Let’s look more closely at the Greenland Ranch reports from 1914.
Portions of the original record form for Greenland Ranch for January 1914 (that are available from NCDC) have apparently been “written over” in an attempt to make the data more legible. (This occurred from time to time in previous months, also. The copy received by the Weather Bureau is a second or third carbon copy, and the data entries are very light and difficult to reproduce.) Unfortunately, it appears that some of the “write-overs” are wrong! For January, 1914, a maximum temperature of 73F is given in the “monthly summary” on the right hand side of the sheet. But, on the 2nd and 4th, the “write-over” maximum temperatures are 86F and 89F. Here are the maximum, minimum, and 5 p.m. “set max” temperatures for the first five days of the month:
1 71/65 68
2 86/65 65
3 71/66 70
4 89/63 67
5 70/64 68
Note that the “set max” temperatures on the 2nd and 4th are well below the maximums that are given. It seems likely that the person who wrote over the original entries mistook a “6” for an “8”, and changed a “66” to an “86” and a “69” to an “89” (and did so without checking the maximum temperature for the month!). Of course, if one looks at the minimum temperatures during this period and if one is familiar with the climate of Death Valley, then a maximum in the 80s on these dates might seem to be a lot more reasonable than a maximum in the 60s. But, we know that the minimums are too warm, so we can’t use that reasoning to “guess” at the maximums. Maximums of 66F on the 2nd and 69F on the 4th are much more reasonable for the first week of January.
On the 2nd, a 5 p.m. “set max” temperature of 65F was observed, which is the low temperature for the day! The extremes for the next day are 71F and 66F, both higher than the previous-day “set max” temperature. Obviously, the observations during this five-day period are flawed on at least three fronts —- the incorrect write-over maximums, the “set max” issue, and the bogus ultra-warm minimums. We can compare Greenland Ranch to reports from (a newly-established station at) Rhyolite, NV, just 25 miles north of Furnace Creek (at 3680 feet elevation). Here are the daily extremes and the 5 p.m. “set max” temperatures from Rhyolite from January 1 to 5, 1914:
1 55/30 50 cloudy (0.02″ precip)
2 56/31 52 clear
3 59/36 54 partly cloudy
4 61/38 61 cloudy
5 60/36 56 partly cloudy
Wind at Rhyolite was from the north each day. The daily ranges here were at typical mid-winter values: in the twenties. The daily ranges of less than ten degrees at Greenland Ranch from the 1st to the 5th suggest continued major issues with the minimum temperature reports there. The average daily range for January 1914 at Rhyolite was 19 degrees, compared to 7 degrees at Greenland Ranch. The average minimum of 61.5F at Greenland Ranch was a mind-boggling 25 degrees warmer than its average minimum for the first 19 Januarys (which is inflated a couple of degrees as it incorporates the unreliable, warm minimums from Jan 1914 and Jan 1915).
There were four days with measurable precipitation at Greenland Ranch in January: 0.20, 0.10, 0.02, and 0.35″. The 0.02″ and 0.35″ entries were the first precipitation amounts since the station opened in June 1911 that included a numeral in the hundredths column that was not a “1” or a “0”!
The average daily range jumped up to 15.6 degrees at Greenland Ranch in February, 1914 (compared to the average of nearly 30 degrees for its first 19 Februarys. The minimum temperatures continued to be unusually warm, averaging 16 degrees above its average from 1912 to 1930. Precipitation amounts this February reverted to old habits, with entries of 0.10, 0.10, and 0.01″. There were a couple of days where the high temperature was lower than the previous day’s “set max”. This month’s NCDC copy is not in the handwriting of Oscar Denton, and does not include the observer’s signature. Perhaps the original copy was in such poor shape that the Weather Bureau made this easily-deciphered replacement.
The sheet provided by NCDC for Greenland Ranch for March 1914 also appears to be a replacement form. Minimums continue to be on the high side, as the average daily range was only 25.7 degrees and the greatest daily range for the month was only 32 degrees. There were three days with the “set max” issue and there was no precipitation. For the five months from November, 1913, through March, 1914, at Greenland Ranch, only two days had a minimum temperature reading lower than 50F! The minimum temperature record for these five months is garbage.
The form for April is in the handwriting of Denton, and is signed by him. As previously stated, a remark on the April 1914 form notes the “split column” on the minimum thermometer. The minimum temperatures are missing for the first three days of the month. There is no information regarding the severity of the “split column”, nor how the problem was remedied. The minimum temperatures and daily ranges for month do not appear unreasonable, except for April 4th (96/83). Presumably, the minimum thermometer was put back in service on the 4th, and the observer recorded the minimum temperature reading even though it was not for a 24-hour period. The average daily range for the month was 26 degrees, compared to the longer-term average of about 30 degrees. The greatest daily range was 45 degrees on April 18 (100/55), though in the monthly summary, the greatest daily range given is 42 degrees. There are, as is typical, a few of the “set max” issues this month. The two precipitation entries at the end of the month are 0.07″ and 0.05″, but these are “write-overs”.
The minimums for May 1914 and for the summer months of 1914 appear reasonable, though the monthly average minimums are all warmer than the averages from 1911 to 1913. For example, the average minimum of 71.3F in May of 1914 was about six degrees warmer than that for the previous two Mays, the average minimum of 76.0F in June of 1914 was warmer than the averages of each of the previous three Junes, and the same goes for July and August. This resulted in somewhat lower average daily ranges during the summer of 1914 — ranges which are much closer to the average daily ranges from the current Death Valley station. Average daily ranges averaged about 30 degrees from May through August 1914. This begs the question — did the ground cover beneath and around the station change compared to earlier years, when the daily ranges were several degrees larger and minimums were cooler? Did cultivation and irrigation of the alfalfa crop cease in 1914, at least in close proximity to the shelter? This may be the case.
Though the temperature data from Greenland Ranch for June, July and August of 1914 do not appear too unreasonable, there are still the occasional “set max” issues. Six times in June, 8 times in July, and five times in August the maximum for the day was lower than the “set max” that was recorded on the previous day. There are no “write-overs”, and the “set max” temperatures have a line through them in July and August. It isn’t clear why the Weather Bureau would “strike out” the “set max” temperatures, as they all appear reasonable with regard to the maximum temperature report for the same day. Apparently the reviewer noticed that some next-day maximums were lower than the “set max” temperature.
On the form for June, 1914, a precipitation amount of “.5” is entered on the 2nd. But, a line has been drawn through this entry, and “.05” is written right next to it. As with many of the precipitation entries up until now while Oscar Denton was the observer, the handwriting appears to be by someone other than Denton. Given all of the previous daily precipitation amounts at Greenland Ranch that are to the nearest tenth-of-an-inch, is it possible that some, or most, of these are incorrectly entered? Did the observer of the rainfall amounts at Greenland Ranch know the difference between “tenths” and “hundredths” of an inch, and did he understand the decimal system so that he could record precipitation properly? On only two additional days during 1914 was precipitation observed and recorded. On both December 12 and December 17, “3/10” is entered, but written on top of that is “.30”.
Big problems arise once again into the autumn and winter months at the Greenland Ranch weather station. Inexplicably, as happened during the same period a year previous, the minimum temperatures do not drop in tandem with the maximum temperatures. By December 1914 and January 1915, the daily range values are consistently at very un-Death Valley-like levels (around 10 degrees). The minimum temperature report for December was 52F, and for January was 50F. These reports are ridiculous — the average daily minimum at Furnace Creek for these two months is in the upper 30s. Let’s take a close look at the last four months of 1914 to see if we can figure out what was going on.
The record for September 1914 at Greenland Ranch appears to be contrived, or artificially manufactured. It seems as though the temperature record was “made up”! The document is signed “OA Denton” and the handwriting is his. Unlike practically all of the months since Denton began as observer, there were no instances of a maximum temperature coming in lower than a previous-day “set max” temperature. Usually, this would be a good thing — perhaps Denton is starting to conform to proper procedure. But, the numbers for September, 1914, are not realistic. The maximums for the month range from 103F to 112F, but all of the maximums from the 1st to the 28th ranged from 107F to 112F. The maximum temperatures from the 12th to the 27th were either 109F or 110F, and the maximum was 110F on nine consecutive days, from the 19th to the 27th! Further, the “set max” temperature was 108F on each of the 14 days with a maximum of 110F. The “set max” temperature was two degrees lower than the maximum temperature report on all but three days.
Minimum temperature reports for this September are more realistic, but still a bit peculiar. They drop from the mid-upper 70s during the first week, to near 70F during the second week, to 60F on the 14th and 15th (giving suspiciously high daily ranges of 50 degrees). Minimums are near 74F for most of the 3rd week, and down to near 66F during the 4th week.
The table below shows the range of maximums from September 1st to September 28th at Greenland Ranch and surrounding stations:
Greenland Ranch 107 to 112F, 5 degrees
Lida, NV 65 to 80F, 15 degrees, including six maximums of 80F, three maximums in the 60s
Las Vegas, NV 80F to 102F, 22 degrees, six maximums of 100-102F, three maximums in the 80s
Barstow, CA 85F to 98F, 13 degrees, seven maximums of 85F to 89F, six maximums from 95F to 98F
Independence 70F to 88F, 18 degrees, 14 maximums from 85F to 88F, five maximums in 70s
(Note: Rhyolite data were missing for September, 1914)
Independence, Lida, and Las Vegas showed a big cool down from the 12th to the 13th (16 degrees cooler at Independence), but Greenland Ranch dropped only two degrees: from 112F to 110F on the maximum and from 110F to 108F on the 5 p.m. “set max”.
The maximums at surrounding stations exhibit typical monthly variations during September 1914, while those at Greenland Ranch suggest an extremely improbable four-week stretch with little change. It appears that the observer did not know the actual maximum temperatures for much, if not all, of September 1914, and that he entered bogus maximums and “set max” temperatures. The minimums do not appear unreasonable. The cool-down around the 13th is reflected in the minimums.
Why would Denton enter bogus values? Why would his superior(s) at the ranch allow bad reports to be mailed to the Weather Bureau? The maximum temperature reports from Greenland Ranch during September 1914 are outrageous and obviously flawed. Perhaps the maximum thermometer was broken, and Denton estimated the maximums. I suspect that the observations were not being taken on a regular basis during this month, and the observer “filled in the blanks” with what he thought were reasonable, believable figures.
Moving on to October, 1914, the average daily range plummets to only 19 degrees, even though the observer characterized all but five of the days as “clear”. Again, it appears that the minimums are staying too high. Maximums in October ranged from 85F to 101F, and the distribution and “look” of the maximums from day to day appear reasonable, unlike those for the previous month. The minimums ranged from 68F to 82F, and appear to be too high right off the bat in October. Here are the daily temperatures from September 29 to October 5, 1914, at Greenland Ranch:
29 105/65 set max 103
30 103/80 set max 101
1 98/82 set max 97
2 97/81 set max 95
3 92/81 set max 90
4 88/75 set max 85
5 91/71 set max 90
There is a curious jump in the minimums from September 29th to 30th, and the minimum temperatures stay uncharacteristically warm thereafter. The daily ranges at Greenland Ranch for the first five days of October 1914 are 16, 16, 11, 13, and 20 degrees. These values are about fifty percent from the average. Given the clear skies indicated at Greenland Ranch, and its location near the bottom of a deep desert basin, daily ranges from 30 to 40 degrees would be expected. Daily ranges at Independence from the 1st to the 5th were 30, 26, 32, 39 and 38 degrees. To the north of Furnace Creek, Rhyolite had daily ranges of 31, 33, 19, 33 and 23 degrees from the 1st to the 5th. Rhyolite is located on elevated terrain, where better mixing keeps daily ranges significantly smaller compared to Greenland Ranch (on average). The very small daily ranges and anomalously warm minimums during the first week of October 1914 at Greenland Ranch again suggest that the minimum thermometer is not functioning properly, and/or that the observer is unaware or is not concerned. The too-high minimums continue into the first few months of 1915, unfortunately.
The average daily range in November 1914 at Greenland Ranch was only 18 degrees, and the average minimum of 63.7F is about 17 degrees above the average for the month. For comparison, at Independence, the average minimum of 32.5F was a degree below average. The “set max” issues continued. There were 8 of these in October and 6 in November. It appears, at least, that Denton was interested in furnishing good maximum temperature data again, despite these breaches of proper procedure!
The handwriting of the entries on the forms for October, November and December of 1914 is that of Denton, but the form for November is signed “F W Corkill”. The “set max” entries have a line through them for each month from July through December of 1914. Again, this was likely drawn by the reviewer of the data at the Weather Bureau, who probably noticed that some next-day maximums were lower than the corresponding “set max” temperatures.
The temperature data set for December 1914 continues to be problematic. Minimums are much too high, averaging 59.4F, more than 20 degrees above average. The average maximum of 72.9F was 8 degrees above the 1911-1930 average. Okay, one might say, perhaps December 1914 was quite a bit warmer than normal, and the data from Greenland Ranch can be trusted. Then consider the first two sentences from Henry F. Alciatore, in his summary for Nevada for December, 1914, in Climatological Data: “Considered as a whole, the month was unusually cold. There were several abnormally cold days in each decade.” The average temperature at Las Vegas (41.8F) was about four degrees below the long-term average, and the average at Independence was a chilly 29.0F, a full ten degrees colder than average! How does Greenland Ranch come in at 14 degrees ABOVE average for this month?
Curiously, the 5 p.m. “set max” temperatures for December, 1914, are closer to the minimum temperatures than the maximum temperatures for much of the month. There are numerous instances where the “set max” temperature is lower than the both the maximum and minimum temperature for the same day and the next day! Impossible!
A cold wave enveloped the region from December 17 to 20. Here are the daily extremes at Independence and Greenland Ranch:
17 Ind 34/9 GR 68/58 set max 50
18 Ind 32/5 GR 69/58 set max 57
19 Ind 25/5 GR 70/56 set max 54
20 Ind 29/3 GR 75/58 set max 54
Where does one begin with this?! The official minimum for the month at Greenland Ranch is 52F, yet the 5 p.m. “set max” temperature on the 17th is 50F! The 5 p.m. “set max” temperatures at Greenland Ranch during these four days more accurately reflect the cold weather that was affecting the region than the maximums and minimums. (Maximums in the 40s and 50s and minimums in the 20s to 30s would be expected at Greenland Ranch during this period, based on the Independence data.) On December 22 and 23, Rhyolite (elevation 3680 feet and only 25 miles north of Greenland Ranch) had daily extremes of 46/18 and 50/19, respectively. Greenland Ranch on these days reported 82/70 and 80/70. Where did that warm air at Greenland Ranch come from? The answer —- it was not that warm there. The temperature data are flawed. It is unwise to consider any of the temperature data from December, 1914, as reliable. Why and how did the temperature data become so poor? The best answer —- the weather station observer did not know how to provide, or did not care about providing, a trustworthy climate record —- and much of the data are “made up.” Perhaps Denton was not at the ranch for part of the month and he “filled in the blanks” later with bogus observations.
Greenland Ranch Data Quality, 1915
Monthly Greenland Ranch forms for 1915:
Monthly coop forms for Independence for 1915 and 1916 are not available from NCDC via the Internet at this time. However, the daily temperatures and precipitation for Independence are listed in California’a Climatological Data issues:
Daily maximum temperature charts for Death Valley and select surrounding stations for July and August, 1915:
The twelve forms for Greenland Ranch for 1915 were signed “OA Denton”. The forms that are available from NCDC for February, March, April, May, October, November and December appear to be “re-makes”. Someone other than Denton, perhaps a Weather Bureau or NCDC employee, wrote over the entries in order to make legible a copy that was faint and difficult to read. The “set max” temperatures were entered by Denton, but are not copied onto the forms for October, November and December.
The precipitation entries continue the trend of recent years, with (apparently!) the amounts rounded off to the nearest tenth of an inch on the wetter days. There were nine days with measurable precipitation at Greenland Ranch during 1915. The first two (on January 28 and 29) were “.50” and “.60″, and the remainder ranged from .01″ to .07”. No precipitation was observed during 1915 after July 23.
As in 1914, the minimums are too high and the daily ranges are too small during the first couple of months. Daily ranges appear to be too low during the summer months, too.
The following table provides the average daily range, the greatest and smallest daily ranges, and the temperature extremes for each month in 1915, at both Greenland Ranch and at Independence. Also provided is the difference in monthly maximum temperature between the two stations for the high-sun months from April to September:
1915 Daily Range Monthly
Ave Gr Sm Mx/Mn
Jan GR 10.5 18 4 76/50
Jan IN 22.5 32 6 56/18
Feb GR 14.4 20 8 75/43
Feb IN 23.2 33 12 60/22
Mar GR 25.6 39 6 92/45
Mar IN 26.7 38 11 78/26
Apr GR 26.6 39 13 101/56 +21
Apr IN 26.7 38 16 80/27
May GR 30.1 42 16 113/52 +21
May IN 31.1 42 17 92/27
Jun GR 27.2 43 9 118/70 +24
Jun IN 31.5 40 23 94/47
Jul GR 18.2 25 8 120/89 +22
Jul IN 32.0 40 24 98/45
Aug GR 23.2 28 14 124/89 +26
Aug IN 35.1 40 27 98/52
Sep GR 29.4 36 11 112/79 +21
Sep IN 30.8 40 21 91/38
Oct GR 28.5 35 22 101/49
Oct IN 36.5 46 24 88/34
Nov GR 16.3 39 10 89/39
Nov IN 30.1 38 22 74/15
Dec GR — 69/mm (20 minimums missing)
Dec IN 25.1 42 11 67/10
The average daily ranges for each month of 1915 are smaller, in some instances much smaller, than the norm for Furnace Creek-area stations for the past 100 years.
The trend of impossibly small daily ranges and strange, warm minimums continues from December 1914 and into January 1915. “Set max” temperatures are occasionally lower than both the maximum and minimum temperature in January 1915, and the minimum for the month is only 52F. Minimums at Independence ranged from 18F to 32F and averaged 23.9F, three degrees below its long-term average through 1930. Minimums in the 30s and 40s would be expected at Greenland Ranch this month. The Greenland Ranch temperature record for this month is entirely unreliable.
Minimums in the 40s show up on the first five days of February 1915 at Greenland Ranch, but then jump inexplicably back up into the 50s for the remainder of the month. The average daily range of only 14.4 degrees is pathetically low compared to the long-term average of about 29 degrees. The “set max” temperatures are suspicious. In January the “set max” temperatures were 5 to 15 degrees lower than the day’s maximum temperature (and sometimes lower than the day’s minimum temperature!). However, in this February, all but one of the “set max” temperatures are two degrees lower than the maximum for the day! The one “rogue” day had a “set max” that was only one degree lower than the maximum. This is similar to September, 1914, when it appeared that most, if not all, of the temperatures had been “made up”. None of the temperature data from February 1915 can be considered trustworthy.
As maximum temperatures crept up into the 70s, 80s and 90s during March 1915, the minimums at Greenland Ranch remained primarily in the 50s and low 60s. The average daily range this month is a more realistic 25.6 degrees, but even that is a bit small for a desert basin station during a month with only one day with rain. The “set max” temperature, again, is two degrees lower than the maximum on EVERY DAY except one! Maximums of 78F at Independence and Rhyolite around March 23-24 would support a maximum in the mid-upper 90s at Greenland Ranch, but the high for the month was only 92F (on the 16th and 23rd). It would not be prudent to trust any of the temperature data from Greenland Ranch this month.
The “set max” temperatures for April 1915 are all one or two degrees lower than the maximum for the date. Thus, considerable doubt is cast upon all of the temperature reports for this month. Eight maximums are crossed out, probably by the Weather Bureau, and the “set max” temperature from the previous day is entered so that the maximum is a true “24-hour maximum.” It might be that Denton was taking observations at different times of the day, and not always at 5 p.m. He subtracted a degree or two from the maximum and entered that value, regardless of the time he read and reset the thermometer. The Weather Bureau data reviewer noted these occasional faulty maximum temperatures that were lower than the previous “set max”, and made the correction. However, these corrections were more the exception than the rule.
The “set max” temperatures become more realistic during the first week of May 1915. The difference between the “set max” and the maximum is only two degrees on the first 3 days of May, but thereafter are mostly 4 to 12 degrees. The average daily range of 30 degrees is close to the norm.
On the 11th, the maximum jumps up to 111F (from 96F on the 10th). This is a bit curious and deserves some scrutiny. The table below shows the maximums at Greenland Ranch and surrounding stations on May 11 and May 31, 1915, when Greenland Ranch reported 113F:
Greenland Ranch 111 and 113
Independence 82 and 92
Rhyolite 78 and 88
Tonopah 69 and 81
Las Vegas 94 and 101
Needles 86 and 104
Barstow 90 and 97
The maximums at the surrounding stations do not support Greenland Ranch’s maximum of 111F on the 11th, but they DO support its 113F on the 31st. Note that the maximums at surrounding stations are much warmer on the 31st compared to the 11th. Maximums on the 11th at the closest stations, 82F at Independence (3907′) and 78F at Rhyolite (3680′), support a high temperature of perhaps 100 to 104F in Death Valley. If the maximum of 94F at Las Vegas is warmed adiabatically to Greenland Ranch’s elevation, then a temperature of about 106F is realized. Tonopah’s maximum of 69F would support a maximum of about 100 to 105F in Death Valley. I would contend that Denton continued to play “fast and loose” with the maximum temperature on some of the hottest days. The “set max” temperature on May 11 was 98F, which corresponds more closely to an afternoon maximum of about 104F (if the maximum thermometer was indeed checked and reset close to 5 p.m.). The report of 111F on May 11th at Greenland Ranch is very questionable.
All of the “set max” entries are two degrees lower than the maximum temperature reports for June 1915,
except for two days when the differences are one and three. The form entries appear to be “write-overs” by the Weather Bureau or NCDC due to a faint original copy. There are eight instances in which the maximum is less than the previous day “set max”, but, unlike for April 1915, the reviewer of the record did not make any changes or corrections. The average daily range of 27 degrees for June 1915 is not unreasonably small (and is only a couple of degrees smaller than latest 30-year normal for Death Valley), but is about ten degrees smaller than the average daily ranges for the first three Junes at Greenland Ranch. Was “cultivated land” still beneath and adjacent to the thermometer shelter at this time? The smaller daily ranges compared to the first few summers would suggest not —- but given the general unreliability of the temperature data during Denton’s tenure, it is difficult to make any definite determinations!
The June 1915 maximum of 118F on June 16th at Greenland Ranch is not well-supported by maximums at surrounding stations (Independence 91F, Barstow 96F, Needles 104F, Rhyolite 94F, Tonopah 77F, and Las Vegas 103F. As on May 11, 1915, the maximum temperature report on June 16th at Greenland Ranch appears to be at least five degrees warmer than the lower atmosphere would be able to muster at the elevation of Furnace Creek, based on these other maximum temperature reports (and the dry adiabatic lapse rate).
Can any of the maximums of June 1915 for Greenland Ranch be trusted? The “set max” reports are bogus. Only five of the maximums are “odd” numbers, and the final 13 maximums are “even” numbers. A quick comparison to the maximums at Independence suggest general consistency with the cooler and warmer periods. Except for the 118F on the 16th, the maximums appear to be okay.
The June minimums appear reasonable — until midmonth. The average daily range for the first 14 days of June 1915 was 31 degrees (109F/78F). The average daily range for the last 16 days of the month was only 17 degrees (112/95). The very small and suspect daily ranges continue into July. The smaller daily ranges beginning on June 15 were due to a sudden increase in minimum temperatures. (June 14 had 105/70, and June 15 had 112/96.) Minimums from June 1 to 14 ranged from 70F to 88F. Minimums for the remainder of June 1915 ranged from 87F to 105F. From the 15th to the 22nd, minimums ranged from 96F to 105F. Summertime minimums that are this warm can and do occur, but generally not in June. They are generally associated with heat waves during July or August. Monsoon moisture and cloudiness is usually part of the weather picture when Death Valley minimums are in the mid-90s or warmer. Did other stations show a big jump in minimums for the last half of the month? Here are the average maximum and minimum temperatures at Greenland Ranch and a few other stations from June 1-14 and from June 15-30:
Gr. Ranch 109/78 112/95
Barstow 90/63 94/65
Rhyolite 88/58 95/63
Las Vegas 97/57 102/60
There is a jump of a few degrees in the minimums at the three other stations, but nothing like the 17-degree jump reported at Greenland Ranch.
Daily extremes on June 17 and 18 were 115/105 and 110/101, respectively. These were the second and third days on record at Greenland Ranch with a minimum as high as 100F. (July 30, 1914, had 115/101.) The minimum of 105F on June 17, 1915, appears unlikely. Here is a sample of the minimums and maximums on the 16th, 17th and 18th (the minimum is listed first here because it almost always occurs before the afternoon maximum):
Gr. Ranch 97/118 105/115 101/110
Independence 59/ 91 54/89 63/86
Lone Pine 60/90 65/92 60/90
Rhyolite 61/94 65/91 65/92
Tonopah 55/77 52/79 53/77
Las Vegas 69/103 62/98 59/94
Needles 62/104 71/106 70/102
Barstow 66/96 64/94 67/90
The maximums at Greenland Ranch during this period appear to be a little on the high side given the surrounding maximums, but may be good. The minimums at other stations in the region are at typical June levels from the 16th to the 18th, and the daily ranges at the basin and valley stations are from about 30 to 40 degrees, as usual. Skies at Barstow were clear on the 16th and 17th and partly cloudy on the 18th, and Needles reported clear skies for all three days. There was no precipitation in the desert regions of California and Nevada during these three days. It appears that the very warm minimum temperatures at Greenland Ranch, if authentic, were not due in part to a humid airmass. The other way that very warm minimums can come about is if winds blow all night and do not permit a surface radiation inversion to form. Typically, though, the low levels of the atmosphere will still cool about 15 to 20 degrees by sunrise from the previous afternoon’s high temperature despite a windy night with excellent mixing.
Given that the minimums of 105F and 101F on June 17 and 18, 1915…
—were not associated with a humid period
—did not occur during an excessively hot period in Death Valley (i.e., with maximums of 120F or more)
—are not supported by surrounding stations (which had temperatures fairly close to average for June)
—and occurred near the beginning of an extended stretch of abnormally (i.e., likely not authentic) high minimums and small daily ranges at the station this summer
…they should be considered very suspect and discarded.
It is not clear, again, as to why the Greenland Ranch minimums suddenly increased midmonth. The other desert stations only showed a minor increase in minimums from the first half of June to the second half. Chances are that the minimum thermometer was not functioning as it should. Perhaps there was another separation in the alcohol, or the indicator was in the wrong part of the thermometer. Occasionally, during very windy conditions, the shelter will shake, or vibrate, and the motion will cause the indicator in a minimum thermometer to “shake down” to a higher value, especially if the thermometer is angled (downward from left to right) a little more from horizontal than it should be. For instance, if the overnight minimum temperature is 80F, and a sandstorm hits around noon while the temperature is 105F, the wind might “shake down” the indicator into the mid 80s or even higher, resulting in a bogus minimum temperature indication when the observer reads the instrument at 5 p.m. But, given the persistent and consistent “too warm” minimums from mid-June and into July, the wind/shake explanation for the two minimums above 100F this month probably does not apply. The numerous problems with the temperature record for June 1915 make it difficult to trust any of the data.
The temperature record for July 1915 at Greenland Ranch is similarly flawed. The “set max” temperature is two degrees lower than the maximum on all days but one. Of course, this strongly implies that the observer was a not afraid to bend the rules! On five maximums, it appears that the data reviewer “wrote-over” the entry by Denton so that it matched the previous-day “set max” temperature. The average daily range of 18 degrees is unusually small for July, and all but one minimum was 90F or higher. Six minimums ranged from 100F to 103F. The average monthly minimum of 96.5F compares to average minimums of 83.5, 79.6, 80.8 and 86.5F for the first four Julys at Greenland Ranch. Yet, the average maximum of 114.7F here in July 1915 is a little cooler than normal for July. A check on the average daily ranges and average maximum and minimums at other stations for July 1915…
Gr. Ranch 18.2 degrees (114.7/96.5)
Lone Pine 39.8 (95.6/55.8)
Independence 32.0 (92.2/60.2)
Barstow 28.9 (98.4/69.5)
Needles 25.3 (103.4/78.1)
Rhyolite 28.3 (96.8/68.5)
Las Vegas 35.4 (101.7/66.3)
…do not suggest an unusually hot or humid July. As with the last 16 days of June 1915, the minimums for July 1915 at Greenland Ranch appear to be at least ten degrees too warm. It is possible that by this time that the shelter was above a drier, more natural ground cover (and not over the cool alfalfa field), which contributed to warmer minimums. But, the persistently narrow daily ranges during this July are too narrow, and cast considerable doubt upon the minimums.
The maximum of 120F for July 1915 at Greenland Ranch is among the lowest of all of the Julys there from 1911 to 1960. Though some of the maximums this month seem a little higher than surrounding maximums would suggest, they appear reasonable for the most part. Given the “too warm” minimums and the bogus and obviously contrived “set max” data, it may be that Denton was not at the station for much of the month, and that he “made up” most of the reports.
All of the “set max” temperatures for August 1915 are two degrees lower than the maximum, except for one! This is very curious and strange, and is the third or fourth time this has occurred. Denton appears to have little regard for the the “set max” temperatures! The average daily range of 23.2 degrees (118.2/95.0) is, again, quite small. On only one date (the 31st) is there a “write over” to match the previous day “set max”. Most amazing is that none of the maximums change by more than two degrees from day-to-day, except from the 10th to the 11th, when the maximum jumps up from 120 to 124F. Is this amazingly “smooth” day-to-day change for real? Perhaps!
The maximums from Lone Pine for August 1915 are also very consistent day-to-day, as only four maximums are more than two degrees different from the previous day. Lone Pine is near Owens Lake in the Owens Valley, about 65 miles west of Furnace Creek, at 3728′ elevation. The 3900-foot difference in elevation compared to Furnace Creek would promote high temperature differences of about 20 degrees on typical hot and sunny summer days. The daily maximum temperature differences between Lone Pine and Greenland Ranch for August 1915, range from 18 to 25 degrees, except for a five-day period from the 10th to the 14th, when the differences were 26, 30, 29, 25, and 27 degrees. Greenland Ranch recorded its hottest maximums of the year during these five days: 120, 124, 122, 120 and 119F. Why did Greenland Ranch maximums seem to jump up about five degrees compared to Lone Pine during these five days? Is this another case where the observer may have been artificially inflating the maximum temperature reports at Greenland Ranch?
Rhyolite is a little under 30 miles north of Furnace Creek, at 3680 feet elevation. It had an average maximum of 99.6F in August 1915, 18.6 degrees cooler than that at Greenland Ranch. The daily temperature differences between Rhyolite and Greenland Ranch were all 21 degrees or smaller, except from the 11th to the 16th, when the differences were 23, 25, 20, 22, 21 and 23 degrees. Comparisons to both Lone Pine and Rhyolite show this subtle “jump” in Greenland Ranch maximums during its week with maximums from 120 to 124F. Can these maximums be trusted? It is not easy to determine, but the bogus “set max” data do not increase confidence levels.
All of the 5 p.m. “set max” temperatures at Greenland Ranch for September 1915 are two degrees cooler than the maximum temperature (for the same day) that was entered by Denton. It suggests that Denton was probably not adhering to the 5 p.m. observation time, and that he was resetting the maximum thermometer in the morning (if not also around 5 p.m.). (Alternatively, it might suggest that Denton was not interested in the “set max” values, and he just decided to subtract two degrees from the maximum and write that figure into the “set max” space just to have something there.) The morning “resets” allowed true daytime maximums, but the practice goes against the official observation time and is not standard observing procedure. Denton should have, at least, made a note of the resets on the forms. Fortunately, by providing extremely suspicious “set max” values, Denton has made it easier to make better assumptions as to the veracity of the temperature data.
On four days during September 1915, a “write-over” maximum has been written in, by the Weather Bureau data reviewer, presumably. The “write-over” value in each instance is the same as the previous-day “set max” and is always higher than Denton’s entry.
The average daily range value of 29.4 degrees for this month seems quite reasonable, until one checks the average maximum (115.8F) and the extreme maximum (112F). The average maximum for the month, as it appears on the observer’s record form, was computed incorrectly. The correct figure is 104.8F. Unfortunately, the mean temperature for the month as it appears on the form (101.1F) was computed from the incorrect average maximum, and the incorrect mean value appears in the September 1915 issue and the 1915 Annual Summary issue of Climatological Data.
(The actual average maximum of 104.8F at Greenland Ranch for September 1915 incorporates the four “write-over” maximums. But since these are based on the “set max” figures, which are obviously untrustworthy, it is probably better to use the original maximums. The 3o maximums as entered by Denton average 104.0F.)
A revised average maximum of 104.0F gives an average daily range of only 17.6 degrees for the month. This is a much smaller daily range than would be expected for any September, especially one dominated by clear to partly cloudy skies. The average daily range was much higher at Rhyolite (27.6 degrees) and at Lone Pine (38.4 degrees), so an average daily range of about 30 degrees at Greenland Ranch would be expected. The Greenland Ranch maximum temperatures this month appear relatively trustworthy, and the minimums are too high, by about 8-14 degrees. The minimum temperatures appeared to jump suddenly on June 15, 1915, and continued too high (seemingly) through September 1915. Uncharacteristically small daily ranges were the result from June through September at Greenland Ranch.
Some important changes in the Greenland Ranch record begin in October 1915. The forms continued to be signed “OA Denton,” but the data entries are not in his handwriting. The forms from October 1915, and through all of 1916, are in the same handwriting as those for March 1914 and for February through June of 1915. (These forms that are not in Denton’s handwriting, were, presumably, filled out by another employee at the ranch.) The 5 p.m. “set max” temperatures are no longer provided beginning in October 1915, and the daily ranges increase to more “reasonable” levels. (During the last seven days of September, 1915, the average daily range at Greenland Ranch was 14 degrees. For the first 7 days of October, 1915, its average daily range was 28 degrees.)
What might have prompted these changes? Did Corkill, or someone else, discover the ridiculously small daily range reports by Denton and the obviously bogus “set max” temperature reports? Did Denton’s superior, Fred Corkill, determine that the observations by Denton were suspect? Did Corkill inspect the Greenland Ranch forms from the summer of 1915 and figure out that Denton’s “set max” temperature reports, nearly all of which were exactly two degrees lower than the day’s maximum temperature, were undermining the credibility of the Greenland Ranch records? Did Corkill decide to monitor the observations by Denton more closely, filling out the forms himself to better ensure data quality? Were the “set max” entries discontinued because the station observer was resetting the maximum thermometer at various, inconsistent times, and/or was not recording the “set max” temperature anyway? Was Denton alone for the most part during the summer of 1915, as usual, and did Corkill return around October 1 to Greenland Ranch to discover the unreliable weather records and a problem with the minimum thermometer? Was a problem with the minimum thermometer remedied around September 30th? Did the U.S. Weather Bureau examine much of the recent Greenland Ranch record, discover the very poor data quality, and contact the station supervisor to try to remedy the situation?
The answers are elusive, but the changes in the Greenland Ranch record from September to October, 1915, seemed to improve data quality for a while.
Despite the (presumed) increased attention to data quality, daily ranges continued to be on the small side at Greenland Ranch during October and November, 1915, and minimums that were too high appear to be the culprit. Minimum temperature reports quit on December 12th, 1915, and do not resume again until February 1, 1916. On the form for December, 1915, is the remark: “minimum thermometer out of order. sent in for repair.” It is likely that the minimum thermometer was not functioning as intended, and/or was not operated and read in a proper fashion by the observer, for the final seven months of 1915.
Though the average daily range at Greenland Ranch in October 1915 was 28.5 degrees (90.0/61.5F), that would still appear to be smaller than expected, given the clear and dry weather during the month and the large daily ranges at other desert basin stations. Average daily ranges at surrounding stations include 36.5 degrees at Independence, 40.9 degrees at Lone Pine, 28.5 degrees at Rhyolite, 39.8 degrees at Barstow, and Las Vegas (42.0 degrees). What happened to those big daily ranges that were common at Greenland Ranch during its first two Octobers (with average daily ranges of 36 and 33 degrees)? Given the small daily ranges before and after this month, and the minimum thermometer repair that was necessary in December, it appears likely that October 1915 was also beset with unreliable minimums. I suspect that the reported minimums were some 4 to 8 degrees too warm in October 1915.
The average daily range for November 1915, according to the official record form for Greenland Ranch, was only 16.3 degrees (71.3/55.0F). But, the arithmetic by the Weather Bureau was incorrect on the minimums, and the average low temperature was actually 49.4F. This gives an average daily range of 21.9 degrees, which is still well below the 1911-1930 average of 29 degrees for November. The average daily ranges for November, 1915, at the closest stations (Rhyolite, 23.3 degrees; Lone Pine, 32.3 degrees; Independence, 30.1 degrees; Barstow, 34.4 degrees, Las Vegas 34.2 degrees) strongly suggest that the Greenland Ranch ranges were too small. Again, the minimums are the likely culprit, and those reported are perhaps 4 to 8 degrees too warm.
As mentioned, the minimum thermometer was “sent in for repair” in December, 1915, and minimums were not reported from December 12th, 1915 through January 31, 1916, at Greenland Ranch. Given that the minimum thermometer needed repair, it is likely that the minimum temperature reports in December are suspect. Most daily ranges from December 1 to 11, 1915, were from 20 to 25 degrees, and the minimums were likely 4 to 8 degrees too warm. The maximums temperatures appear reasonable and trustworthy for this December. A very cold air mass covered the region late in the month. Maximum temperatures on the 29th, 30th and 31st:
Greenland Ranch 45, 43, 48F
Rhyolite 44, 29, 35F
Lone Pine 55, 57, 53F
Independence 32, 24, 34F
Barstow 60, 38, 42F
Las Vegas 45, 43, 38F
Tonopah 19, 18, 21F
(The maximums at Lone Pine appear to be too high and are suspect.)
Greenland Ranch Data Quality, 1916
Monthly Greenland Ranch forms for 1916:
Monthly forms for Independence are not available for 1916, except for December. The daily temperature data can be found in the California Climatological Data issues, however.
CA CD 1916 01
CA CD 1916 02
CA CD 1916 03
CA CD 1916 04
CA CD 1916 05
CA CD 1916 06
CA CD 1916 07
CA CD 1916 08
CA CD 1916 09
CA CD 1916 10
CA CD 1916 11
CA CD 1916 12 Inde 1916 12
Daily maximum temperature charts for Death Valley and select surrounding stations for July 1916:
The twelve “cooperative observers’ meteorological record” forms for Greenland Ranch for 1916 were signed “OA Denton”, but the entries are not in his handwriting. It seems likely that Denton’s superior at the ranch filled out the official forms with the observations that were taken by Denton. The “set max” temperatures are no longer provided, which makes it more difficult to determine which maximums might be from the previous-day thermometer reset at or around 5 p.m. It also makes it more difficult to “quality control” the maximum temperatures, and to determine how careful the observer was and whether or not the data set is trustworthy.
The minimum thermometer was not in service for January, 1916, and minimum temperatures resume on February 1, 1916. Minimums are also missing from June 9-17, 1916, but there is no remark given as to why. There were five months with measurable rain at Greenland Ranch in 1916, and eight days total with measurable rain entries. Three of the daily amounts were “.01”, and the other five were to the tenth of an inch.
The following table provides the average daily range, the greatest and smallest daily ranges, and the temperature extremes for each month in 1916, at both Greenland Ranch and at Independence. Also provided is the difference in monthly maximum temperature between the two stations for the high-sun months from April to September:
1916 Daily Range Monthly
Ave Gr Sm Mx/Mn
Jan GR —– 67/m (minimums missing)
Jan IN 21.3 32 8 52/-2
Feb GR 24.4 33 16 82/32
Feb IN 25.0 30 20 60/2
Mar GR 26.1 43 10 98/46
Mar IN 33.6 44 21 80/26
Apr GR 22.2 30 13 104/61 +20
Apr IN 34.5 41 26 84/33
May GR 31.4 42 21 110/57 +24
May IN 35.0 44 17 86/28
Jun GR 31.3 45 22 121/67 +24 (9 mins missing)
Jun IN 36.5 48 29 97/40
Jul GR 31.6 42 20 127/76 +24
Jul IN 38.0 47 28 103/45
Aug GR 29.3 42 20 121/75 +25
Aug IN 33.4 42 22 96/48
Sep GR 36.3 45 25 116/63 +26
Sep IN 35.2 46 22 90/38
Oct GR 29.1 39 17 98/44
Oct IN 30.8 41 18 78/26
Nov GR 31.2 38 17 80/30
Nov IN 33.8 43 22 73/16
Dec GR 27.8 36 15 73/24
Dec IN 29.5 46 13 60/-2
The daily range values (provided above) for Greenland Ranch during 1916 look okay and appear reasonable at first glance. The averages are relatively close to the average daily range values for the past 30-40 years at the “Death Valley” weather station in Furnace Creek (i.e., near 30 degrees June-Nov and near 28 degrees Dec-May). Compare these numbers for 1916 versus the ridiculously low average daily ranges from the winters of 1913-14 and 1914-15, and versus the consistently smallish daily ranges for the final six months of 1915. Note that daily ranges of less than 20 degrees are relatively rare, and that all of the daily ranges from May through December in 1916 were 15 degrees or greater. Also, the greatest daily ranges for each month are quite consistent, and generally from 35 to 45 degrees for most months.
January 1916 was one of the coldest, cloudiest and wettest months on record for the region. (The Greenland Ranch minimums were missing, unfortunately!) The district forecaster, G.H. Willson, began his general summary of the month with “The weather of January, 1916, was the most remarkable experienced in California in many years.” At 6090 feet elevation, Tonopah, NV, had maximums as cold as 13F to 18F on four days. Lone Pine received 48 inches of snow, and 70.5 inches of snow fell at Independence! The average high at Independence was only 37.3F, and at Greenland Ranch was 56.5F (about 9 degrees colder than average). Denton measured 1.50″ in the gage at Greenland Ranch midmonth. The coldest maximums at Greenland Ranch were 47 and 45F on the 13th and 14th. Rhyolite showed highs of 37 and 34F on these two days. The maximums at Greenland Ranch appear to be trustworthy.
The minimum temperature reports returned in February. The daily ranges are a little smaller than normal in February, March and April. There are a couple of suspiciously warm minimums of 75F on March 18 and 19 (with maximums of 88 and 90F). Breezy and/or very cloudy conditions are typically responsible for allowing very warm minimums on occasion at Furnace Creek. The observers’ notes at Rhyolite and Greenland Ranch reveal that is was partly cloudy to cloudy, with moderate south wind on the 19th after “calm” on the 18th.
The greatest daily range of only 30 degrees for April is suspiciously small. There were plenty of days with clear skies at Greenland Ranch (25). To its north, Rhyolite, a better-exposed station with smaller daily ranges on average, ten days had daily ranges of 30 to 35 degrees in April of 1916. Its average daily range this month of 27.4 degrees is significantly greater than Greenland Ranch’s 22.2 degrees. Are the old problems of high minimums occurring again?
The maximums and minimums at Greenland Ranch look good, at a glance, from May through July. A major heat wave during the last two weeks of July pushed the maximum temperature above 120F on nine days, including 127F on the 23rd and 126F on the 24th. Between these two very hot afternoons was an overnight minimum of only 103F on the 24th.
There is a stretch with suspicious temperatures at Greenland Ranch from June 27th through July 7. On these 11 consecutive days, the maximum temperature was an “even” number. (Back in September, 1914, there was an unlikely stretch of 21 days, 18 of which were “even” numbers and nine straight 110F highs.) Was Denton “making up” maximum temperatures again, entering “guesses” on days with no observer, no observations? Eight of the eleven minimums from June 27 to July 7 were “even” numbers. Greenland Ranch maximums during the eleven days averaged 27 degrees warmer compared to Rhyolite and 24 degrees warmer than maximums at Independence. The difference in elevation between Furnace Creek and these other two stations supports a temperature difference of about 20 degrees. On July 2nd, Greenland Ranch’s maximum of 114F appears much too warm compared to the 76F at Rhyolite, 85F at Independence, 88F at Barstow, 94F at Las Vegas, and 69F at Tonopah. Unfortunately, as long as Denton is the observer, it is ill-advised to blindly accept as trustworthy ANY data from Greenland Ranch. Fortunately, the maximums of 120F to 127F during July, 1916, are supported by the temperature reports from surrounding stations.
August 1916 maximums and minimums at Greenland Ranch appear reasonable, as do those for the first half of September 1916, but then the “attack of the 110s” occurs again. From September 17 to 29, the maximums range from 109 to 114F, including three 109s and eight 110s. The daily temperatures from Greenland Ranch from September 21 – 29, 1916:
21 110/74
22 110/74
23 109/72
24 110/66
25 110/66
26 110/72
27 110/75
28 109/64
29 109/65
It was during the same time frame two years earlier that Greenland Ranch had very similar reports! Here are the daily temperatures from September 19-28, 1914:
19 110/74
20 110/74
21 110/74
22 110/67
23 110/67
24 110/66
25 110/66
26 110/65
27 110/65
And, a check of maximums at surrounding stations throws much doubt onto the maximums at Greenland Ranch from the 21st to the 29th. While all Greenland Ranch maximums ranged from only 109 to 110F, maximums at Rhyolite ranged from 67 to 80F, were from 74 to 84F at Independence, 78 to 92F at Barstow, 61 to 75F at Tonopah, and 86 to 97F at Las Vegas. How did the dubious Greenland Ranch maximums come about during this period of September 1916? The observer likely was not at the ranch, and he “guessed” both the high and low temperatures when he returned and wrote those on his form. Apparently the person who filled out the official record for Denton (i.e., Corkill?) was unaware of the dubious entries; or he was aware, and still permitted the bad data to be sent to the Weather Bureau.
Of course, the average maximum and minimum temperatures for September 1916, and the average daily temperature range, are compromised if Denton “estimated” temperatures for a week or more . It would probably be prudent to throw the entire month out rather than to allow it to figure into any Greenland Ranch long-term averages.
The temperature data for Greenland Ranch for October, November and December of 1916 appear reasonable. A stretch of two weeks from December 8 to 22 was characterized by clear skies and light winds for the most part over the Death Valley region. The dry and stable air mass allowed for typically moderate to large daily ranges at Greenland Ranch. For these 15 days, the average maximum was 60.0F, and the average minimum was 27.4F. Maximums ranged from 58 to 63F, and minimums were all below freezing, ranging from 24 to 30F at Greenland Ranch. These temperatures might seem to be quite cold for Death Valley, for a station below sea level during a fair weather period with sunny days.
This is a good opportunity to show how temperatures behave in the region during periods of nice, fair mid-winter weather. Clear skies and light winds during the winter result in relatively cold conditions in the lower portions of desert basins. The short days, low sun angles, and long nights beneath clear skies cause a deficit of radiation at the surface. Whereas in summer the lower atmosphere in these desert basins is active and well-mixed, the opposite is common in winter during fair weather. Light wind at night promotes fairly strong radiation inversions in and near the basin bottoms. Midday sunshine is strong enough to break these relatively shallow inversions, and to allow a little mixing with warmer air at higher levels; but insolation is much too weak to generate the deep afternoon convection to heights of 5,000 to 14,000 feet above ground level that is commonplace in summer. The long nights can promote a general “building” of cold air in the basins, as the heavy and dense nature of cold air is difficult to dislodge or mix out. Relatively warm air masses can spread over the basins, thousands of feet above, but have little effect on temperature in the basin bottoms.
During patterns such as these, some of the higher and better-exposed stations can be just as warm, or warmer, than the low-elevation stations at or near the basin bottoms. A weather station at Rhyolite operated from late 1913 to early 1917, and was the closest station to Greenland Ranch until Cow Creek opened in the 1930s. The Rhyolite station was at 3680 feet elevation and about 25 miles north of Furnace Creek. As mentioned, the average maximum/minimum at Greenland Ranch from December 8-22, 1916 was 60F/27F. For the same period, Rhyolite averaged 54F/29F. Even though Rhyolite is nearly 4000 feet higher than Greenland Ranch, the minimums there averaged slightly warmer. In summer, maximums at Rhylolite average about 20 degrees cooler than at Furnace Creek. During these 15 mid-winter days, it was only 6 degrees cooler. Since Rhyolite is situated on a slope and is not at the bottom of a deep basin, the air cooled near the surface at night does not linger as it does in the bottom of Death valley. The cold air at Rhyloite mixes out more easily with warmer air aloft. For the nine days from December 13 to 21, 1916, Greenland Ranch (60F/27F) and Rhyolite (57F/29F) averages were quite similar. Farther south, at Indio (elevation -20 feet), this 9-day period averaged 71F/31F. Though Greenland Ranch and Indio are at similar elevations, the mid-winter maximums are very different. Sunny winter afternoons in Coachella Valley are much warmer than in Death Valley primarily because of topography. The Coachella Valley is not a narrow and deep basin (or a “bowl”, of sorts), where heavy, cold air is difficult to flush out. The lower latitude of the Coachella Valley and its slightly higher sun angles on winter days would also help promote warmer maximums, but the “lay of the land” is the most important determinant.
In 1916 at Greenland Ranch, the minimum thermometer was repaired and returned to service on February 1. The average daily ranges for the year appear to be much more reasonable and reliable compared to the very suspect and small daily ranges for parts of 1913, 1914 and 1915. There are several periods during 1916 with suspicious temperature data. But, fortunately, a repeat of the very bad, warm minimums of the winters of 1913-1914 and 1914-1915 did not occur. The average daily ranges of 1916, however, were still generally smaller than those of 1911 and 1912. Why is this the case? If we can assume that the instruments were functioning properly, then there might have been a major change in the ground cover beneath the station’s thermometer shelter sometime from late 1913 to late 1915. The smaller daily ranges in 1916 versus 1911 and 1912 are likely due primarily to warmer minimums, because the station was no longer directly above an irrigated, cultivated field. There is no official record of a station move or a change in ground cover.
(Note—Temperature data for Lone Pine and Needles during 1916 appear to be very suspect at times.)
Greenland Ranch Data Quality, 1917
Monthly Greenland Ranch forms for 1917:
Monthly Independence forms for 1917:
Daily maximum temperature charts for Death Valley and select surrounding stations for July 1917:
The Greenland Ranch record forms for 1917 were signed by Oscar Denton. However, the data entries for January through October are in the same handwriting as those for 1916 — not Denton’s, but perhaps by Denton’s supervisor. The November and December forms are in Denton’s handwriting. The “set max” temperatures are not provided in 1917. There were only ten days with measurable rain entries this year. Only one was more than 0.09″: “.30″ on May 24th. On eight days .01 to .03” was observed and recorded, and on January 1, an amount of “.001″ was entered. The Weather Bureau counted that amount as .01”.
The following table provides the average daily range, the greatest and smallest daily ranges, and the temperature extremes for each month in 1917, at both Greenland Ranch and at Independence. Also provided is the difference in monthly maximum temperature between the two stations for the high-sun months from April to September:
1917 Daily Range Monthly
Ave Gr Sm Mx/Mn
Jan GR 25.7 35 11 66/24
Jan IN 42.2 57 27 72/-1
Feb GR 27.4 39 20 66/18
Feb IN 30.0 44 15 66/18
Mar GR 30.7 42 15 96/36
Mar IN 31.8 43 16 76/20
Apr GR 33.8 42 21 105/43 +21
Apr IN 32.6 41 22 84/21
May GR 30.1 40 25 106/51 +20
May IN 31.0 38 22 86/34
Jun GR 33.8 46 21 123/66 +25
Jun IN 35.3 44 19 98/44
Jul GR 29.3 44 19 125/78 +23
Jul IN 33.1 43 26 102/54
Aug GR 35.9 45 24 123/71 +25
Aug IN 34.9 41 22 98/51
Sep GR 31.4 4 18 115/61 +21
Sep IN 36.7 46 25 94/43
Oct GR 32.5 40 13 106/44
Oct IN 37.7 44 30 92/31
Nov GR 30.1 38 19 87/35
Nov IN 34.5 47 22 78/25
Dec GR 30.2 36 22 74/28
Dec IN 33.1 40 17 68/18
A quick glance at the numbers above reveals no obvious and prolonged problem with the temperature reports in 1917. Average daily ranges of 30 to 35 degrees are typical. The somewhat small range of 25.7 degrees for January might be due to more cloudiness than normal. The greatest daily ranges are not too far from 40 degrees, also very typical for Furnace Creek. The cloudiest and windiest days sometimes have daily ranges from only 10 to 20 degrees, and these are the smallest ranges reported during 1917. We’ll look a little more closely at the Greenland Ranch temperatures this year to see if Denton provided a reliable set of data.
January, 1917, was “the coldest January weather experienced in Nevada in 28 years,” according to the summary for Nevada in Climatological Data for the month. Greenland Ranch had a mean temperature of 46.4F (59.3F/33.6F), about six degrees colder than the long-term average (1912-1960). The coldest period, around January 15-19, was characterized by clear skies and strong north winds (15-17) followed by cloudier and calmer weather (18-19). Maximums from January 15 to 19:
Greenland Ranch 53/48/49/50/51
Rhyolite 28/24/25/37/39
Lone Pine 32/32/30/35/37
Barstow 44/42/40/37/41
Las Vegas 44/37/44/53/48
Tonopah 8/ 4/ 7/25/33
The maximum of 4F at Tonopah on January 16 is among the coldest on record there (three maximums of -5F, -3F, and 0F occurred in January, 1937). Strong winds, bright sunshine, and excellent low-level mixing caused maximum temperatures in the region to change with elevation at rates close to the dry adiabatic lapse rate (about 5.4 degrees/1000 feet) from the 15th to the 17th. Note that on the 18th and 19th, as warmer air aloft advected into the region and surface winds relaxed, maximums warmed 15-25 degrees at well-exposed stations (such as Tonopah and Rhyolite) while the basin stations warmed little. (Note: the observer at Independence switched from the Weather Bureau to a private citizen late in 1916, and a significant increase in daily ranges and maximum temperatures on sunny days suggests a change to a location which was not as well exposed. Maximum temperature reports at Independence from the 15th to the 19th were 38, 32, 36, 36, 56.)
The temperature data for Greenland Ranch for the first few months of 1917 appear to be problem-free. A warm spell in late April caused maximums over 100F in Death Valley, and the Greenland Ranch maximums stay suspiciously warm for a day or two as surrounding stations cool significantly. Here are the maximums from April 24-28, 1917:
Greenland Ranch 103/104/104/105/85
Independence 92/ 94/ 78/ 66/ 60
Barstow 90/ 90/ 77/ 75/ 81
Las Vegas 92/ 95/ 90/ 78/ 73
Tonopah 69/ 71/ 54/ 49/ 38
The time of observation is around 5 p.m. at these stations, except for Tonopah (midnight-to-midnight). It is possible that some of the maximums from the 26th to 28th reflect conditions on the previous afternoon, but the maximum of 105F at Greenland Ranch on the 27th must be bogus if the surrounding maximums can be trusted. (Notes: Rhyolite closed in March, 1917; Lone Pine maximums are very suspect for April and May, 1917, as all maximums were from 68F to 81F for the two months. Maximums at Independence ranged from 50F to 94F in April and from 62F to 86F in May.)
June, 1917, was hot, and Greenland Ranch’s six maximums from 121F to 123F are supported by maximums at the closest desert stations.
The average maximum temperature at Greenland Ranch for July, 1917, was 121.9F. This is the highest average maximum temperature for any month for any Death Valley weather station, except for Badwater. All of the maximums ranged from 119F to 125F. It is the only month for Furnace Creek-area stations in which all maximums were 119F or hotter. Most minimums were in the 90s, two minimums were 100F and one was 105F. Unfortunately, the maximums temperatures during the final week of the month are suspect. Did the observer estimate some temperatures for several days?
The summary for Nevada for July, 1917, states that the month was one of the warmest Julys on record in Nevada up until that time, with temperatures “uniformly high during the entire month, the daily averages being above normal with the exception of a day or two on or near the 28th when they were normal or slightly lower.” Also, “an unusually warm spell occurred from about the 10th to the 15th when very high maximum temperatures were recorded.” The daily maximums at Greenland Ranch from the 1st to about the 23rd appear trustworthy, but the maximums change little from the 24th through the 31st while maximums at surrounding stations take a pronounced dip (see Figure below).
A comparison of average maximums for the two periods shows a decrease of 6 to 11 degrees region wide, except at Greenland Ranch:
Average maximums from July 1-23, 1917 and July 24-31, 1917, with difference:
Greenland Ranch 121.7/122.3/+0.4
Oasis Ranch, NV 98.0/ 88.9/ -9.1
Tonopah 90.8/ 83.4/ -7.4
Las Vegas 107.4/ 96.3/-11.1
Needles 111.8/104.1/ -7.7
Barstow 104.9/ 98.3/ -6.6
Lone Pine 98.0/ 92.5/ -5.5
Independence 98.4/ 91.4/ -7.0
From July 27 to 29, surrounding stations reported maximums which were 5 to 10 degrees or more below normal as moist air, showers and thunderstorms impacted the region. Rainfall was measured at Greenland Ranch on the 27th and 29th, with 0.03″ and 0.02″, respectively. Despite the clouds, moisture, and lower-than-normal maximums at the closest stations, the maximums of 123F, 120F, and 123F at Greenland Ranch for the 27th, 28th and 29th were well above the long-term July average of 115 or 116F. The minimum temperatures at Greenland Ranch are reminiscent of previous periods during the summer at Greenland Ranch —- periods with curious reports which raise a red flag of caution. From July 5 to July 30, 1917, all of the minimums are even numbers, except for four minimums of 95F and one of 105F. (For comparison, eight minimums ended with a “1”, “3”, “7” or “9” in June 1917 and eleven did in August, 1917.) Given the suspicious minimums at Greenland Ranch and the late-month hot temperature reports (despite the region-wide cool-down), can ANY of the maximums for July 1917 be trusted? The amazingly persistent maximums for the month, all ranging from 119 to 125F (with 29 of 120F-plus!) might be valid through about the 21st, but should be disregarded thereafter. The (record) average maximum of 121.9F is not legitimate. I suspect that Denton was not at the ranch for an extended stretch of time or two during July 1917, and he recorded his best “guesses” for the maximum and minimum temperatures for a number of days.
The dubious data for Greenland Ranch for July 1917 extended into August, 1917, as the run of consecutive 120F-plus maximum temperatures continued from July 6 to August 17. All of the high temperature reports for this 43-day stretch ranged from 120F to 125F. As discussed above, several maximum temperature reports of 120F or more at Greenland Ranch are likely too high during the last week of July, 1917, and a similar scenario plays out during the first half of August, 1917. Greenland Ranch maximums range from 120F to 123F for the first 17 days of August, 1917, according to Denton’s reports. However, a significant cool-down affected the region around August 11. Barstow had a maximum of 106F on the 9th and only 90F on the 11th. Independence showed maximums of 98F on August 5 and 6, and maximums of 90F each day from the 11th to the 14th. Las Vegas had maximums of 106F to 108F from August 6-10, and maximums of only 93F to 99F from August 11-14. Similarly, Tonopah maximums dropped from the upper 80s into the mid 70s, and Oasis Ranch maximums were as high as 97F on the 10th before tumbling to 80F and 82F on the 11th and 12th. Showers and thunderstorms were associated with the cooler weather which began around the 11th. How does one explain the maximums at Greenland Ranch? They ranged from 120F to 123F from August 1 to 10, were still as high as 122F on both the 11th and 12th, and 120F to 122F through the 17th. Why didn’t maximums decrease in Death Valley around the 11th and 12th?
Once again, the minimums at Greenland Ranch suggest that daily observations were not made by Denton (or anyone else) for a part, or for parts, of August, 1917. From August 7 to August 15, all of the minimums were even numbers, except for one that ended in a “5” (similar to July, 1917). This was the period with the region-wide cool-down, when Greenland Ranch maximums changed little and remained at or above 120F. I suspect that Denton again “guessed”, or “made up” several, if not many, of the maximum and minimum temperatures for Greenland Ranch during the first half of August, 1917. I suspect that daily observations were missed for most of the first half of August, and probably for most of July, 1917, too. The weather was indeed extremely hot during most of this period, and I suspect that Denton abandoned the ranch for a week or so at a time to escape the heat, and he wrote in “guesses” for the daily temperature extremes based on the thermometer readings upon his return(s). The “character of the day” reports for Greenland Ranch appear accurate enough (Denton provides either “clear” or “cloudy”), which suggests that he may have been hanging out in the cooler air of the nearby mountains, while watching the sky and perhaps making weather notes.
The last two weeks of temperature reports for Greenland Ranch in August, 1917, appear more realistic and may be trustworthy. The monthly maximum of 123F for August, 1917, is likely valid, as is the minimum of 71F. However, the dubious nature of the temperature data from July and through mid-August casts a shadow over all of the reports this summer. The average daily extremes for July and August are not trustworthy, and the user of the daily data would be well advised to consider every entry as suspect.
Recall previous Septembers at Greenland Ranch (since Denton began as observer) with curious temperature patterns, such as many consecutive maximums of “110” and long periods with just even-numbered minimums. It was postulated that the observer was absent during these periods, and made-up “best-guesses” found their way onto the official forms. The maximums and minimums for Greenland Ranch for September and October, 1917, are occasionally inconsistent when casually compared to surrounding stations. From September 28 to October 9, all of the minimums are again either even numbers or end in a “5”. From the 28th to the 5th, all of the maximums range only three degrees, from 103F to 106F, while the surrounding stations had much wider ranges of maximums. This matches the occasional pattern of relatively consistent maximums and even-numbered or “5-ending” minimums, all of which are likely “guesses” by the observer for periods when he was not at the station. After the nine consecutive “even or 5” minimums through October 9, about half of the remaining 22 minimums for the month were “odd”.
A quick perusal of the record for Greenland Ranch for November and December, 1917 does not indicate any suspicious or problematic temperatures.
Greenland Ranch Data Quality, 1918
Monthly Greenland Ranch forms for 1918:
Again, the monthly forms for Independence are not available online. The California Climatological Data forms are provided here, which provide the daily temperature data for Independence.
All of the Greenland Ranch observer forms for 1918 are signed by Oscar Denton, but only those for January and February are in Denton’s handwriting. It is postulated, or assumed, that Fred Corkill or another supervisor filled out the remaining forms that were to be mailed in to the Weather Bureau, using the observational data provided by Denton. The “set max” temperatures are not entered, and the daily temperature range values are provided on the forms for September through December.
There were nine days with measurable rain reports at Greenland Ranch during 1918. Denton entered “3/10″ for Feb 2, March had five entries (.20, .05, .30, .10 and .10″), April shows a day with .05”, and the remainder of the year had only two traces and two days with “.01”. These reports continue the common theme of precipitation entries since the station opened, with most measurements to the nearest tenth of an inch or just one hundredth of an inch.
The following table provides the average daily range, the greatest and smallest daily ranges, and the temperature extremes for each month in 1918, at both Greenland Ranch and at Independence. Also provided is the difference in monthly maximum temperature between the two stations for the high-sun months from April to September:
1918 Daily Range Monthly
Ave Gr Sm Mx/Mn
Jan GR 29.1 38 19 74/30
Jan IN 26.5 39 13 71/20
Feb GR 26.6 39 15 77/30
Feb IN 25.1 39 8 71/14
Mar GR 22.5 34 5 91/40
Mar IN 23.9 36 6 77/27
Apr GR 28.3 38 16 97/50 +16
Apr IN 28.3 38 14 81/27
May GR 26.0 40 13 108/48 +20
May IN 28.9 40 4 88/34
Jun GR 27.2 42 9 124/68 +23
Jun IN 29.2 38 18 101/52
Jul GR 24.0 44 5 120/75 +23
Jul IN 30.0 37 20 97/53
Aug GR 26.7 41 15 125/72 +25
Aug IN 30.5 38 22 100/46
Sep GR 31.7 48 15 112/54 +21
Sep IN 28.6 36 12 91/47
Oct GR 26.0 39 12 100/40
Oct IN 27.6 36 14 84/32
Nov GR 26.6 42 14 85/30
Nov IN 25.3 35 10 75/23
Dec GR 23.5 34 12 66/21
Dec IN 23.2 36 10 66/18
Average daily ranges by month for the prior year, 1917, at Greenland Ranch ran from about 28 to 36 degrees, except for the 26 degrees for January. The average daily ranges for the twelve months of 1918 take a considerable plunge, ranging from only 23 degrees to 29 degrees, except for the 32 degrees for September. This raises a big red flag, as average daily ranges tend to stay fairly consistent from year to year at stations with reliable records. Sure, some months might be a lot cloudier and wetter and/or windier than average which would result in significantly lower average daily ranges. But, this persistent and significant decrease in average daily ranges for all of the months of 1918 as compared to 1917 strongly suggests that the temperature reports may be problematic —- for one year or the other. There is the possibility that the station and the instruments were moved, perhaps to a better-exposed location, but that is not indicated in any of the records. There is also the possibility that there was no “cultivated ground” or irrigation near the station during this year, which could contribute to warmer minimums.
Usually, when a large and significant shift in the daily ranges is noted (without a move or a change in local ground cover), it is due to a problem with one (if not both) of the thermometers. A common problem with the standard Weather Bureau alcohol minimum thermometers in use at this time was a slow sublimation of alcohol into the reservoir at the top of the thermometer. This would result in a gradual lowering of the minimum thermometer readings, and the issue might not be noticed in a timely manner if the reset readings of the maximum and minimum thermometers are not checked regularly. For example, minimum temperature indications off of a minimum thermometer with this issue might read 1, 2 or 3 degrees lower than what it should read. The substation record for Greenland Ranch states that this type of problem was rectified several times during the 1940s and 1950s, as separations of several degrees were remedied.
The average daily range for all of 1918 was 26.5 degrees, compared to 31.7 degrees for 1917. A five-degree change one year to the next is very significant and unusual. Recall that the minimum temperatures were missing for all of January, 1916, and that the minimum thermometer was “out of service for three days on account of a split column” in April, 1914. The minimum thermometer already had a history of problems, so perhaps another split column issue, or “evaporation and collection of alcohol into the reservoir” issue was remedied at the station during the winter of 1917-18, resulting in more accurate (and warmer) minimum temperature reports at Greenland Ranch during 1918. This would result in smaller daily ranges.
By the way, the average daily range for 1918 is very close to the average daily range of 27.1 degrees for the later “Death Valley” station at the Visitors Center in Furnace Creek for the 30-year normal period from 1961-1990. If the daily range for 1918 is reliable, then it suggests that the station was no longer impacted heavily by evaporative cooling from surrounding (or nearby) agriculture and irrigation. This is a tremendous decrease compared to the average daily ranges for the first couple of years of record at Greenland Ranch. There is a big caveat, however. The 1918 data may be plagued by plenty of estimated values, giving the resulting dubious averages, extremes, and ranges.
The temperature data for January and February appear credible. March was a rather wet month in the Death Valley region, and the five days with rain at Greenland Ranch match up well with the wet days at Lone Pine and Las Vegas (where monthly totals of 2.80″ and 2.63″ were recorded). The numerous cloudy days and storm systems held the average daily range at Greenland Ranch to a moderate 24.5 degrees. However, the dreaded “evens” and “5s” appear again on the March form. These curious and unexplainable stretches where all or most of the maximum and minimum temperatures are even numbers or factors of “5” may be during periods when the observer, Oscar Denton, is not available to make the daily observation. It is postulated that Denton estimated the daily temperatures for these periods. Also common during these stretches are relatively steady maximums and minimums. For the first 14 days of March, 1918, all of the maximums and minimums were even numbers or ended in a “5”. From February 26 to March 22, only 3 of the 50 temperature values ended in “1”, “3”, “7” or “9”. In fact, 32 of the 50 values are perfectly divisible by 5! More believable temperature reports are entered for March 23 to 31, as only 3 of the 18 values end in a “5” or a “0”, and six of the 18 end in “1”, “3”, “7” or “9”. Compared to surrounding stations, the maximums for Greenland Ranch for the first two or three weeks of March 1918 match up somewhat poorly on a handful of dates, but surprisingly well on many. Denton may have been guessing the maximum and minimum temperatures for much of this March, but since Greenland Ranch is so far from the nearest stations, and much lower in elevation than most, it is difficult to know for certain just what to believe. Any user of the record from February 26 to March 22 should view the data with a lot of skepticism.
The “evens” and “5s” return in early April. The first 9 days show maximums and minimums of this type, except for one minimum of 51F. Of the 30 total maximums and minimums for the last 16 days of April (data for the 30th are missing), 13 end in a “0” (60, 70, 80 or 90F), 24 are even numbers, and only 6 are odd. What is going on? Why is the 30th missing? Did Denton lose a day during April? Can any of the data be trusted? Some of it look okay, some appear suspect.
The temperatures for May appear a little more reasonable, though 13 of the 62 entries end in “0”. The day-to-day temperature trends match up reasonably well with Independence and Barstow, and daily ranges seem to be a little on the low side during the second half of the month.
June 1918 was marked by two heat waves, with five maximums of 121 to 124F. These are supported by maximums at surrounding stations. Statistically, about 20 percent of the minimums and 20 percent of the maximums (or six of each) would be expected to be evenly divisible by “5”. However, 13 maximums and 13 minimums end in either a “5” or a “0”. Denton may have been away for a time or two this month and guessed at a handful of temperatures.
July 1918 is even worse in this regard. Twenty-two of the maximums end in a “5” or a “0”, as do 12 minimums (compared to a statistically likely number of 6 or 7). Only two of the minimums end in a “1”, “3”, “7” or “9”. The relatively cool monthly maximum of 120F and the relatively cool average maximum of 110.1F are supported by surrounding stations. If Denton was not around for many of the observations, and if he indeed was guessing or estimating some maximums and minimums, then it seems as if he was incorporating the occasional readings that he did make as guides and many of his estimates were in the right ballpark, so to speak. So, the monthly maximum is likely valid, and the average maximum might be within a degree or two of reality, but the day-to-day numbers cannot be fully trusted. The statistically unlikely abundance of “0” and “5” values from time to time this month, and this year, strongly suggest a suspect temperature record. A dark pall of suspicion looms over each and every report by Denton, but at least his preference for certain numbers allows the researcher to know which data may require extra scrutiny, or skepticism.
Among the minimum temperature reports for July, 1918, is the all-time record “highest minimum” of 110F, on the 5th. Overnight minimums of greater than 105F in the Furnace Creek area of Death Valley are very, very rare. Minimums of 100F or warmer are typically associated with maximums well into the 120s, windy nights, and sometimes an increase in humidity and cloudiness. Is the 110F minimum authentic? Below are listed the daily extremes (min/max/min/max/min/max) for Death Valley region stations from July 4 to July 6, 1918:
Greenland Ranch 90/120/110/120/88/116 (clear all three days)
Las Vegas 72/103/ 62/105/ 68/106 (cdy/cdy/cdy; trace on 4th and 5th; 6 p.m. “set max” readings of 90/91/92)
Needles 81/116/ 74/113/ 79/ 112 (cdy/pc/pc; trace on 4th; 0.06″ on 5th)
Barstow 70/102/ 71/103/ 72/105 (clr/clr/clr)
Independence 67/ 95/ 63/ 96/ 59/ 95
Oasis Ranch 58/ 96/ 65/ 98/ 58/ 91
Goldfield, NV 57/ 91/ 57/ 92/ 58/ 93 (clr/clr/clr; trace on 4th)
Beatty, NV 63/105/ 63/104/ 58/103 (clr/clr/clr; prevailing wind S/W/S)
A heavy thunderstorm was reported at Goldfield between midnight and 1 a.m. on the 4th. Though clear skies were reported at Greenland Ranch, Barstow, Goldfield and nearby Beatty for the 4th, 5th and 6th, showers and clouds impacted areas not too far to the east and southeast of Death Valley during this period. An increase in humidity and cloudiness could have helped to keep temperatures up to some extent at Furnace Creek on the night of the 4th and early on the 5th, when Greenland Ranch temperature reportedly did not drop below 110F. However, none of the minimums for this overnight period at surrounding stations are especially warm, nor are they any warmer than the nights before and after, except for Oasis Ranch. The closest station to Furnace Creek this July is Beatty, Nevada, about 30 miles to the northeast of Greenland Ranch. The maximums at Beatty (elevation 3309 feet) support the 120F maximums at Greenland Ranch, but its minimum of 63F on the 5th does not support a 110F minimum in Death Valley. The 30 to 40-degree daily ranges at these stations are typical for summer in the Mojave Desert region. The tiny daily range of 10 degrees at Greenland Ranch for the 5th is difficult to explain, and difficult to believe. The air mass was hot, but not especially so for July. Maximums were a few degrees above average for July. For the air to remain at 110F or above for the entire night in the thermometer shelter at Greenland Ranch, an especially persistent and strong wind would have had to blow all night long, and perhaps a mid-level cloud deck would have to be covering the basin, too. Given the typically large daily ranges at all of the surrounding stations, the generally low quality of the Greenland Ranch record while Denton was the observer, AND the curiously large number of seemingly estimated maximums and minimums that are factors of “10” this summer, AND since Death Valley minimums that are warmer than 105F are extremely infrequent; it would be ill-founded to regard the minimum of 110F on July 5, 1918 as authentic.
Perhaps even more “incredible” than the minimum of 110F on the 5th, is the maximum of 85F on the 8th. Is it possible that the extremes of only 85F and 80F on the 8th at Greenland Ranch are valid, given the following reports for the 8th?
Gr. Ranch 85/80 0.00″ clear
Beatty 98/70 0.00 partly cloudy
Las Vegas 97/67 0.01″ cloudy, shower at 2 p.m., “set max” temp of 79F at 6 p.m.
Barstow 97/66 0.06″ cloudy, shower between noon and 1 p.m.
Independence 94/66
Goldfield 86/51 0.70″ partly cloudy, rain from 4:30 p.m. to 9 p.m., “set max” temp of 59F at 7:30 p.m.
Oasis Ranch 94/57 0.00″
(note: none of the maximums shown above appear to be “previous-day/set-max” maximums)
There were afternoon showers and storms in the region, but it is inconceivable that the maximum at Greenland Ranch would be cooler than at all of these other higher-elevation places. Denton recorded July 8 as “clear” at the ranch, and the extremes for July 7 (108/88F) are both warmer than the maximum that he entered for the 8th. The maximum temperature observation of 85F for July 8 at Greenland Ranch is almost certainly bogus.
Another period with clouds and showers occurred at the end of July, 1918, and the daily extremes of 100F and 92F for July 31st at Greenland Ranch are suspect.
August, 1918, commences with a heat wave, with 125F and 123F at Greenland Ranch on the 3rd and 4th. These values appear a little inflated, but not entirely unreasonable, given the maximums at Independence (99 and 100F), Goldfield (94 and 95F), Beatty (106 and 104F), Las Vegas (108 and 104F), Barstow (105 and 103F on the 2nd and 3rd), Oasis Ranch (97 and 98F) and Tonopah (90 and 91F). The 11 maximums from the 5th to the 15th are comprised of seven “110s”, a “95”, a “105”, a “115” and a “108”. The “0” and “5” values are back, suggesting that Denton once again was estimating some daily temperatures. The values for the last half of the month look better than those for the first half, and may be good.
September also got off to a warm start, but the observation of 110F/95F for September 2 may be estimated. A quick comparison of maximums versus those at Beatty and Independence show general consistency, but with occasional outliers at Greenland Ranch that are probably guesses or estimations. Of the first 19 maximums for the month, ten fall into the “0” or “5” bin, as do ten minimums. Only four or five of each would be expected statistically.
It is unfortunate, but the first order of business when trying to determine the veracity of the temperature data at Greenland Ranch during 1918 is to count the number of entries which end in “0” and “5”. On average, about 20 percent, or 12 of the 60 or 62 temperatures should end in a “0” or a “5”. For October, 1918, 13 maximums and 15 minimums are multiples of “5”. The total (28 out of 62) is more than double that which would be expected statistically. A region-wide warm spell from October 10-13 is not reflected in the maximums at Greenland Ranch, where seven of the eight temperatures end in a “0” or a “5”. It must be assumed that Denton was not at the ranch during this period, or simply was not doing his job, and was estimating many temperature observations.
The first six maximums for November, 1918, are: 75, 85, 85, 75, 65 and 65F. Fifteen maximums this month and fourteen minimums end in a “0” or a “5”. Ugh. User beware.
The number of “0” and “5” entries for December decrease to somewhat more reasonable totals of 7 maximums and 12 minimums. The monthly low of 21F near the end of December looks okay.
The temperature data for Greenland Ranch for 1918 can not be trusted. It appears that a large percentage — 20 percent? 30 percent or more? — of the entries for the year may be compromised, may be estimates by the observer. Some estimates appear reasonable and may be based on good readings off of the thermometers, which may have gone unread for days or weeks at a time. The “estimates” are conservative for the most part and may have resulted in the rather low-to-moderate average daily ranges from month to month. The maximums and minimums for each month of 1918 are probably reliable, but the daily data in general should be considered suspect, especially during timeframes with an inordinate number of “0” entries, “5” entries, and/or even-numbered entries.
Greenland Ranch Data Quality, 1919
Monthly Greenland Ranch forms for 1919:
The Climatological Data issues for California for 1919:
The twelve meteorological record forms for Greenland Ranch for 1919 were signed “OA Denton”. Those from June to November are in Denton’s handwriting. “Set max” temperatures are not provided. Only four days showed measurable rain amounts, two with 0.01″, 0.20″ on November 27 and 0.30″ on December 4.
The following table provides the average daily range, the greatest and smallest daily ranges, and the temperature extremes for each month in 1919, at both Greenland Ranch and at Independence. Also provided is the difference in monthly maximum temperature between the two stations for the high-sun months from April to September:
1919 Daily Range Monthly
Ave Gr Sm Mx/Mn
Jan GR 29.7 38 16 73/19
Jan IN 28.5 37 18 66/16
Feb GR 29.3 44 12 77/29
Feb IN 21.7 30 12 65/21
Mar GR 31.1 45 11 90/30
Mar IN 26.0 37 12 73/23
Apr GR 32.7 45 22 104/46 +18
Apr IN 29.9 40 18 86/33
May GR 31.4 41 23 112/60 +16
May IN 28.5 38 15 96/49
Jun GR 36.3 50 29 121/55 +24
Jun IN 33.9 41 19 97/48
Jul GR 28.9 40 18 122/78 +23
Jul IN 30.9 36 25 99/54
Aug GR 27.2 40 15 123/80 +22
Aug IN 32.4 39 23 101/52
Sep GR 25.7 41 17 111/65 +20
Sep IN 29.9 39 14 91/41
Oct GR 25.8 37 16 93/43
Oct IN 29.8 39 21 77/24
Nov GR 27.3 39 13 83/30
Nov IN 29.2 38 7 71/18
Dec GR 26.3 35 12 68/22
Dec IN 23.0 30 12 62/18
The average daily ranges by month for 1919 are not unreasonable, which suggests (at least) no obvious, persistent problematic data during the year. There is a noticeable decrease in daily ranges from the first half of the year to the second half.
The maximums and minimums from January to March appear reliable for the most part, but towards the end of March and into April the “evens” and “5s” dominate, suggesting again that daily observations were missed and Denton was estimating many readings. From the two weeks from March 31 to April 13, only one of the total of 28 maximums and minimums ended in a “1”, “3”, “7”, or “9”. The day-to-day maximums during these two weeks do not “jive” particularly well with those at two of the closest stations, Beatty and Independence. For the last 17 days of April, 12 of the 34 maximums and minimums end in a “1”, “3”, “7” or “9”, which is close to that which would be expected statistically. The apparent “estimations” during the first two weeks of April by Denton appear decent enough, or close enough, to provide relatively adequate average maximums and minimums for April. However, it is not possible to trust the day-to-day reports by Denton. Unfortunately, none of the temperature reports for April, 1919, can be fully trusted, and it would be ill-advised to utilize these values in long-term data sets to determine “record maximum” or “record minimum”, and the like, for the date. And, obviously, if the observer is not actually observing for a number of days during a month and is estimating daily extremes, then values such as the “lowest maximum” and “highest minimum” for the month may not be reliable.
The record for May and June follows similarly. The temperature data do not appear unreasonable in a casual comparison to surrounding stations. The extreme maximums and minimums for the months are probably valid. But, from May 10 to June 30, more than 50 percent (27 out of 51) maximums end in a “5” or a “0”. The statistically likely total would be only about ten, or 20 percent. It appears that Denton estimated a fairly sizable number of daily temperatures during this period, based on his penchant for using certain numbers.
On Greenland Ranch’s form for July, 8 maximums and 14 minimums end in a “5” or a “0”. The data look reasonable, though, compared to Independence, Barstow, Las Vegas and Beatty. Were some temperatures estimated? Who knows?! The same goes for August, 1919. There are 11 maximums and 16 minimums that end in a “5” or a “0”. For a month with 30 or 31 days, only 6 or 7 maximums and 6 or 7 minimums of this type would be expected on average. These “0” and “5” values continue to dominate the daily listings during the warm season, and this suggests that Denton was likely taking refuge from the heat from time to time in the nearby higher elevations, and estimating daily temperatures for Greenland Ranch for several days at a time based on the readings off of the instruments upon his return.
Previous Septembers have been plagued with dubious temperature reports at Greenland Ranch, and September, 1919, continues that trend, at least during the first week. Eleven of the first 16 reports (maximums and minimums from September 1-8) fall into the “0” and “5” bin. In addition, the daily ranges in large part appear to be too low for this time of the year. Despite 29 out of 30 clear days and near-to-greater-than-average daily ranges at surrounding stations, the average daily range of only 26 degrees at Greenland Ranch is a full four degrees less than average. This, in large part, is likely due to many poorly-estimated temperatures.
The somewhat disturbing trend of somewhat sub-normal daily ranges continued into October. The average daily range was only 26 degrees once again, compared to long-term October averages closer to 30 degrees for Furnace Creek stations. The number of “0” and “5” reports was only 15 for October, which is down appreciably and much more reasonable, statistically, compared to the previous 4 or 5 months. This suggests that Denton was making regular observations. Most Octobers have at least one or two days with a daily range well into the 40s, but this October’s greatest daily range was only 37 degrees. The reason for the smallish daily ranges is not apparent. It may be that the minimum thermometer may have had a separation and may have been reading a few degrees too high, but this is speculation.
The preponderance of “0” and “5” reports returns in November, 1919. Ten maximums and 16 minimums are reports of this kind. A bitter and cold storm system on the 27th gave Greenland Ranch 2/10 inch of rain and a maximum temperature of only 45F. (But, is this figure trustworthy? Maximums on the 26th and 27th were 60F and 45F, and minimums were 40 and 30F. Was Denton estimating these?) Other impressively chilly maximums on the 27th include 25F at Independence and 12F at Tonopah. Maximums on the 27th at Beatty (62F) and Pahrump (52F) do not support a maximum of only 45F at Greenland Ranch, but these reports may be previous-day “set max” values. The max of 44F at Las Vegas is more supportive. When observation time is close to sunset, as was the case at most cooperative stations during the first few decades of the 20th century, comparisons of high temperatures on the coldest days is not as straight-forward as on the hottest days because of the possibility of previous-day, leftover “set max” situations. Most stations did not provide “set max” information, unfortunately.
The Greenland Ranch temperature reports for December, 1919, do not appear unreasonable.
Greenland Ranch Data Quality, 1920
Monthly Greenland Ranch forms for 1920:
The Climatological Data issues for California for 1920:
This is the final full year with Oscar Denton as observer for Greenland Ranch weather station. (The last monthly record form signed “OA Denton” was February, 1921. Victor Ceballos replaced Denton beginning March, 1921.) The forms for January through June, 1920, are signed “OA Denton”, but only April and May are in his handwriting. The forms for January, February and March were filled out by someone else, perhaps by Denton’s superior at the ranch. (Since some of the NCDC forms in Denton’s handwriting are faint carbon copies and difficult to read, it may be that a more legible copy was made at Greenland Ranch, or by another Pacific Coast Borax employee at Ryan or Death Valley Junction, for the Weather Bureau.) The form for July is marked as “copy” in the remarks section. In this instance, it appears that the copy was made by the Weather Bureau or NCDC, probably because the original copy from Greenland Ranch was difficult to read and photocopy. The August form is faint and in Denton’s handwriting through August 20. The remaining observations for the month look to be in the handwriting of Victor Ceballos, and the form is signed “JBoyd”. The September form is not signed and is in Ceballos’ handwriting. The October form is in the handwriting of Ceballos through the 10th, and in Denton’s handwriting for the remaining 21 days. It is signed “OA Denton.” The forms for November and December are by Denton’s hand, but are unsigned. (Note: a “Major Julian Boyd” was listed in the 1921 “Directory of the American Institute of Mining and Metallurgical Engineers, Inc.” Boyd was an Australian mining engineer, a goldminer, an instructor on goldmining at the University of Southern California, and a former Australian soldier, injured during World War I. Boyd was hired around 1920 by the Pacific Coast Borax Company as superintendent of mining operations at Ryan. As postmaster at Ryan, he handled the mail and the monthly weather reports from Greenland Ranch.) “Set max” temperatures are not provided during 1920.
Daily precipitation amounts in 1920 follow the familiar pattern for Greenland Ranch, now in its tenth year as a cooperative weather station. All nine rainfall measurements…or should I say “reports”, were to the “tenth of an inch.” Six of the amounts were 0.10, 0.20, or 0.30″. The three remaining amounts were 0.60, 0.60, and 0.70″. Why were all of the amounts to the nearest tenth of an inch this year, and for much of the previous nine years? Who knows?! Why didn’t anyone during the first ten years at Greenland Ranch know how to measure rain amounts to the nearest hundredth of an inch? Why didn’t the Weather Bureau notice this trend and make sure the observers were trained properly?
The following table provides the average daily range, the greatest and smallest daily ranges, and the temperature extremes for each month in 1920, at both Greenland Ranch and at Independence. Also provided is the difference in monthly maximum temperature between the two stations for the high-sun months from April to September:
1920 Daily Range Monthly
Ave Gr Sm Mx/Mn
Jan GR 23.0 35 10 70/32
Jan IN 24.4 31 10 71/24
Feb GR 21.2 31 10 74/32
Feb IN 22.6 33 5 64/27
Mar GR 21.8 37 8 82/41
Mar IN 25.9 34 15 72/25
Apr GR 25.5 40 10 101/46
Apr IN 27.9 37 18 82/30
May GR 29.9 41 19 114/56
May IN 29.7 36 23 94/39
Jun GR 25.9 39 16 120/60
Jun IN 29.0 36 14 97/46
Jul GR 28.4 37 15 125/71
Jul IN 33.6 39 22 100/54
Aug GR 25.9 39 8 124/82
Aug IN 30.9 37 16 124/82
Sep GR 25.6 40 9 112/57
Sep IN 30.9 42 22 93/39
Oct GR 21.6 40 10 96/45
Oct IN 28.3 39 12 85/28
Nov GR 22.4 34 13 80/33
Nov IN 19.4 35 10 69/23
Dec GR 25.5 35 15 71/28
Dec IN 25.5 38 12 69/21
The monthly average daily ranges above are on the low side for Greenland Ranch. Average daily ranges which run persistently small (as seems to be the case during 1920) or persistently large usually suggests an instrument problem or some other reporting or recording issues. As mentioned previously, changes in instrument location, exposure and/or ground cover can and do affect daily temperature range data significantly. One problem that the researcher has with regard to the early Greenland Ranch record is that the first several years the thermometer shelter was above “cultivated land”, an irrigated alfalfa field, which would contribute to quite sizable daily ranges. By the 1920s, images of the station show that the shelter is above bare ground, immediately adjacent to cultivated land and later seemingly well-removed from cultivated land. It is not known exactly when the changeover occurred, and it is possible that the ground cover changed intermittently during the year and from year-to-year during the first 10 to 15 years of record.
The average daily range in 1912 was 35.7 degrees. This was the first full calendar year of operation, during which the instruments were presumably operating properly, were not in need of repair or replacement, and during which the new observer(s) appeared to be following procedures properly. The observing and equipment problems and the apparent change in ground cover from about 1913 to 1920 make it difficult to determine what a “normal” average annual daily range actually is at Greenland Ranch. Is the 35.7 degree average daily range in 1912 reasonably close to “normal” for a shelter above a well-irrigated alfalfa field at Greenland Ranch? Is the average daily range of 24.7 degrees for 1920 near “normal” for a shelter at Greenland Ranch which is NOT affected by any cooling affects due to agriculture? The average daily range at nearby Cow Creek for its 27 years of operation was about 27 degrees. Compared to Greenland Ranch, Cow Creek was at a slightly higher elevation, was a bit better exposed, was not near irrigated land, and was likely a little less impacted by shallow nighttime surface inversions during light winds. (Additionally, the temperature records from Cow Creek are trustworthy.) The average daily range at “Death Valley” from 1961 to 1990 was near 27 degrees, and that has increased to 28.5 degrees for 1981-2010. The increase in daily ranges in recent decades appears to be due to a gradual decrease in exposure to wind as vegetation increases in breadth and height around the visitor’s center and thermometer shelter, resulting in warmer maximums.
The average daily range of 24.7 degrees for 1920 would seem to be abnormally small for a shelter above bare ground at Greenland Ranch. The years just prior to 1920 at Greenland Ranch appear to be plagued by conservative temperature “estimates” that would reduce the average daily range by a couple of degrees or so. Let’s see if that trend continues for 1920.
The temperature data for Greenland Ranch for January, 1920, may be compromised some. From January 22 to February 1, the maximums and minimums are in the notorious “evens and 5’s” mode. During these 11 days, only one temperature report ended in a “1”, a “3”, a “7” or a “9”. Denton may have missed observations for about ten days during this period, and estimated the temperature observations. A warm spell during the last week of January, 1920, bumped the temperature up to 55F at Tonopah, 78F at Las Vegas, 63F at Independence, 84F at Beatty, and 74F at Pahrump, and 66F at Barstow. But, Greenland Ranch could only manage a 70F. The problem with comparing maximums in winter, though, is that there may be significant differences with regard to low-level mixing from station-to-station. Cold air in deep basins that is not mixed out can result in maximums there that are significantly cooler than at higher elevation stations where wind and mixing has eliminated any local cold-air pooling. Without more details on wind and sky cover at Greenland Ranch, it can be ill-advised to simply point at any particular mid-winter maximum and to deem it suspect just because it seems too cool compared to maximums at higher elevations. I would mark the last ten days of temperature reports during January, 1920, as “suspect” given the preponderance of “evens and 5’s” and the poor track record of data quality during the Denton years.
Though the maximums and minimum temperature reports for February do not appear to be unreasonable when compared to those at nearby stations, there again are an unusual number of “5” reports (12 of the first 40 maximums and minimums end in a “5”, when only 4 would be expected statistically) and even-numbered reports (the last ten minimums for the month are even). This suggests that Denton was estimating many observations. As an example, here are the daily extremes at Greenland Ranch from the 7th to the 10th, when cold air and precipitation affected the region:
7) 65/45
8) 65/45
9) 44/32
10) 55/45 0.70″
and at Independence, Beatty, and Pahrump:
Inde. Beatty Pahrump
7) 60/32 56/27 71/27
8) 48/34 56/32 56/35
9) 38/29 46/30 49/39
10) 49/32 50/28 45/36
The first thing that jumps out are the “double-5″ reports on three of the four days at Greenland Ranch. I suspect that the observer was not at the ranch and did not take the observations on the weekend of February 7/8, 1920. Denton likely returned around the 10th, and estimated the daily extremes for each of these four days. At the surrounding stations, rain was measured on both the 9th and the 10th, so the 0.70” report at Greenland Ranch on the 10th likely includes rain that fell on both days. The maximum of only 44F at Greenland Ranch on the 9th is cooler than at Beatty (32 miles to the north at 3309′) and Pahrump (52 miles to the ESE at 2668′) and is probably an estimate. It appears to be some 5 to 10 degrees too cool, and is even cooler than both the maximum and minimum for the previous day. The reports from the 7th to the 10th from Greenland Ranch may be reasonably decent estimates by and large, but they cannot be considered absolutely trustworthy.
The average daily range of only 22 degrees at Greenland Ranch for March 1920 is suspiciously small given the average daily ranges for the month of 32 degrees at Beatty, 26 degrees at Barstow and Independence, and 31 degrees at Las Vegas. All of these stations are near the bottom of their respective river valley or basin, and average daily ranges near 30 degrees would be expected during early spring. Of the 31 maximums at Greenland Ranch, ten end in “0” (60, 70 or 80F). And, the associated minimum temperature is an even number on nine of these ten days. On the 27th, the extremes of 62F and 54F at Greenland Ranch are not supported by those at Beatty (56/22), Pahrump (51/29), Barstow (55/32), Las Vegas (64/35), and Independence (58/25). Why was the daily range only 8 degrees at Greenland Ranch on this clear day, while the closest stations had temperature ranges of 22 to 34 degrees?
The same trend continues into April. The average daily range of 25.5 degrees for the month at Greenland Ranch is a little on the small side. For comparison, the average daily range was 28 degrees at Independence, 36 degrees at Beatty, and 31 degrees at Pahrump and Barstow. Nine maximums and nine minimums ended in a “0” at Greenland ranch in April 1920. What are the odds that all of these are legitimate observations? (Answer: minimal!) The maximum of 101F on the 30th at Greenland Ranch is well-supported by the 92F at Barstow, 85F at Beatty, 89F at Pahrump, 94F at Las Vegas and 82F at Independence.
The number of “0” temperature reports diminishes some in May 1920 (8 maximums and 3 minimums), and the average daily range is a healthier 30 degrees (though still a tad small comparably: 30 degrees at Independence, 37 at Barstow, 39 at Beatty, 34 at Pahrump and 37 degrees at Las Vegas). Perhaps Denton was around for more actual observations and not estimating as many daily extremes. An impressive springtime heat wave developed mid-month and pushed the mercury up to 114F at Greenland Ranch on the 19th. Again, this temperature is well-supported by the maximums at surrounding stations (101F at Barstow, 96F at Beatty, 101F at Pahrump, 104F at Las Vegas and 94F at Independence).
During the previous several summers and for a number of months leading up to June 1920, the observer’s record at Greenland Ranch suggested strongly that Denton was often absent and “making up” (or, more generously: “estimating”) the daily maximum and minimum temperatures. Fortunately, he based his estimates on the instrument readings upon his return, and he generally had a good handle on how the temperature and weather had behaved in the area during his absence. It is probable that Denton was occasionally spending time in the less-lonely and (usually) more comfortable mining town of Ryan, fifteen miles southeast of Furnace Creek and 3200 feet higher. Ryan was developed beginning in 1914.
Fortunately for the researcher, Denton was not especially imaginative when it came to the figures he used when he estimated maximums and minimums. He often used even numbers, and often used numbers evenly divisible by “5”. This makes the task of identifying potentially or likely suspect temperatures a lot easier. Given the 15 maximums and 10 minimums that end in a “0” or a “5” on the June form for Greenland Ranch, Denton apparently estimated a good number of daily temperatures this month. The inordinate amount of “0” and “5” reports continued into July: 19 maximums and 15 minimums; and for August: 8 maximums and 10 minimums through the 20th. Victor Cebellos recorded the observations on the last 11 days of August 1920, with a total of five “0” and “5” temperatures. Without a diligent, day-by-day comparison of the Greenland Ranch reports with the temperature records at surrounding stations, one cannot be certain that any particular Greenland Ranch report is trustworthy during this summer. The good news is that Denton’s estimates appear to be “in the right ballpark”, and the extreme maximum and the extreme minimum for each month are likely reliable. The average maximum and minimum values during months with many “estimated” observations might be a degree or two more conservative than what would otherwise be observed. Based on Denton’s reports, the resulting average maximum would be a degree or two lower and the average minimum a degree or two higher.
In June, 1920, a wet and cool regime affected the region, and rain was measured at the area stations on both the 27th and 28th. Greenland Ranch reported no rain on the 27th and 0.60″ on the 28th, but that is probably a 2-day total. The maximum of 95F on the 27th and 85F on the 28th are very cool for late June, and the “5” ending suggests that Denton may have guessed on the maximums, but the surrounding station maximums do support the cool maximums.
Given all of the apparent estimated figures this summer, the low temperature of 105F on July 11th can not be trusted. The high of 125F on July 31 is a little warmer than what might be expected based on the closest surrounding maximums (99F at Independence, 107F at Barstow, 102F at Beatty, 107F at Pahrump and 109F at Las Vegas), but it is not unreasonably warm. Tonopah had a maximum of 91F on the 31st: air warmed adiabatically from Tonopah’s elevation of 6090 feet to Greenland Ranch’s elevation of 168 feet below sea level will see an increase of temperature of about 34 degrees, which is the exact difference in observed maximums on the 31st. During the week-long stretch of “record” maximums at Greenland Ranch in July 1913, including three days with highs of 130F or more, Tonopah reached 91F on three days and 92F on two days. It had 89F on the 10th, when Greenland Ranch reported 134F. It should go without saying —- the ultra-high maximums at Greenland Ranch in July 1913 are not supported by the maximum temperature reports from Tonopah.
On the Greenland Ranch record form for August, 1920, the first 20 days of record were entered by Denton, and the 11 remaining were by Victor Cebbalos. Following the high of 125F on July 31, Denton entered extremes of 115F and 106F for August 1, with cloudy skies. The entry for Beatty by E.E. Palmer on August 1 (100F/62F, clear skies) offers little support for the 106F minimum at Greenland Ranch, and that for the more distant Las Vegas (by C.P. Squires) offers scant support (100F/83F, cloudy skies). The report of 101F/87F at Beatty on August 2nd would definitely support the low temperature of 106F at Greenland Ranch on the 1st…except that it is not on the 1st! Maybe one of the two observers was off by a day.
Of course, comparing minimums is much more hazardous compared to maximums, as nighttime inversions of varying strength can cause large temperature differences over short distances and from basin to basin.
The extremes on the 8th and 9th at Greenland Ranch were 120F/101F and 118F/106F, respectively (both marked as “clear”). Beatty’s report of 100F/72F on the 9th, is not entirely unsupportive of the very warm overnight minimum of 106F in Death Valley. If the wind blows all night at Greenland Ranch, and if it does not blow all night at Beatty (along the Amargosa River), a difference of 34 degrees between the two minimums is plausible. The elevation difference of about 3500 feet between Furnace Creek and Beatty would account for about 19 degrees of temperature difference due to compression. Still, a drop of only 14 degrees at Greenland Ranch (from the previous afternoon’s maximum of 120F) during this dry and clear period in the desert is uncharacteristically small.
Can the 106F minimums this month be trusted? Given the general unreliability of the observer; the (presumably) estimated temperatures this month based on the preponderance of “evens and 5’s”; and given that legitimate overnight minimums warmer than about 104F in the Furnace Creek area are extremely rare, I would label the two 106F minimums in August, 1920, as “very suspect”, and I would exclude them from any climatological summary or compilation.
Interestingly, a rare summer cool spell invaded Death Valley during the final 10 days of August, 1920, when Ceballos was the observer. Minimums fell from the 80s and 90s into the 70s immediately upon Ceballos’ arrival. Maximums ranged from only 86F to 95F from the 24th to the 27th when skies were cloudy, and all stations in the region observed precipitation.
The Greenland Ranch form for September, 1920, appears to be in the handwriting of Cebellos, and is not signed. The average daily extremes were 98.3F and 72.1F, but the Weather Bureau computed the average minimum as “39.1F” (well below the monthly minimum of 57F). The sum of the minimums shown on the form is about 1000 less than the actual sum of the minimums. The bad monthly average of 68.4F is published in Climatological Data, September 1920, and presumably the incorrect average minimum and monthly mean are incorporated into subsequent climate tables and averages for Greenland Ranch. A notable cool spell during the last week of the month gave Greenland Ranch two consecutive maximums of only 80F. The maximum and minimum reports by Ceballos appear good this month, though there are no “set max” temperatures provided. These often help in the detection of problematic data.
The entries on the Greenland Ranch form changed from the handwriting of Ceballos to Denton on October 10th. The entries return to the “0” and “5” mode that Denton prefers. Seven of the first ten daily temperature extremes by Denton, from the 10th to the 14th, end in a “0” or a “5”. For the 22 maximums entered by Denton, half are multiples of “5”. For the minimums, 12 of 22 are multiples of “5”. Statistically, only one in five of each would be expected on average or only 4 or 5 out of 22.
Daily data by Denton for November, 1920, do not appear unreasonable. The “0” and “5” regime returns in December, 1920, however. From the 6th to the 31st, 17 of the 26 maximums and 13 of the 26 minimums are evenly divisible by “5”. The plethora of apparently estimated temperature data suggest that Denton was not at Greenland Ranch around the holiday season, or he was simply not taking the observations as tasked. None of the data for December, 1920, can be trusted.
The Denton Years at Greenland Ranch Come To An End
Monthly Greenland Ranch forms for 1921:
California CD’s for 1921
The Greenland Ranch record forms for the first two months of 1921 were filled out by Denton. The January form is signed by Boyd, and the February form by Denton. Thereafter for 1921, the forms are signed by, and are in the handwriting of, Victor Ceballos. Based on the record, Ceballos became the new observer at Greenland Ranch on March 1, 1921. The last entry by Denton was on February 28, 1921. In Climatological Data, “O.A. Denton” was shown as observer for Greenland Ranch through 1921, and “Pacific Coast Borax Co.” was listed as observer beginning in January, 1922. (Though Denton was the observer beginning about September, 1912, Climatological Data showed “J.W. Corkhill” as observer from 1912 through 1914. “O.A. Denton” was labelled as the observer in issues from 1915 to 1921.) Ceballos worked for the Pacific Coast Borax Company at Furnace Creek, and was the weather observer at Greenland Ranch from 1921 to 1927.
The observation by Denton for New Years Day, 1921, was 65F/45F. Was Denton actually at the station on this day to read the thermometers? The “5” entries suggest not. For January, 1921, 17 of the 31 maximums end in a “5” or a “0”. The bulk of these run in consecutive stretches, presumably when Denton did not read the instruments daily and “estimated” the daily temperatures. On the minimums, 10 of the 31 entries end in a “5” or a “0”. Three days with rain amounts are recorded this January: 1/10, 2/10, and 1/10.
February 1921, was without rain, 15 of the 28 maximums were in the “5” or “0” category, as were 7 of the 28 minimums.
This study is concerned first and foremost with the veracity of the record maximum temperatures at Greenland Ranch in July 1913, when Oscar Denton was the station observer. Thus, it is very useful to seek to determine the character, reliability, trustworthiness, dependability, competence and integrity of Denton. Was Denton someone who could be trusted to provide quality climate data for Greenland Ranch all of the time, on a consistent basis? The answer would seem to be a resounding “no”, based on the frequent curious, suspect, and unsupportable data sets during Denton’s seven-plus years as observer.
The next question might be “Why?” Why was Denton an unreliable observer? Of course, we will never know the best answer with any certainty without more details on his background, education, etc. It would appear that Denton was ill-suited for the job on several levels. He likely did not have a scientific background, and he likely did not have a strong appreciation for the importance of providing high quality and trustworthy data. Denton appeared to have no qualms with regard to “bending the rules” when it came to time of observation and when to reset the thermometers. He disregarded previous-day “set max” temperatures and reported maximums that were lower then the temperature was on the “reset”. Perhaps Denton was not adequately trained or properly briefed with regard to taking observations properly, and what to do during certain situations. Denton appeared to have no qualms when it came to “filling in the blanks” when observations were missed. The data suggest that he frequently “made up”, or “estimated” the daily extremes, up to a week at a time. Much of this is conjecture, but how else to explain the unexplainable? How and why do observations keep coming in as even numbers, or ending in a “5” or a “0”, much more than mere chance would dictate, time and time again? Additionally, Denton never seemed to get the hang of observing and recording rainfall amounts to the nearest hundredth of an inch, at least when amounts were more than a tenth of an inch.
There were a couple of occasions during Denton’s time as observer when the temperature observations were obviously faulty and unrepresentative of actual conditions. These were due to an instrument issue in some instances. Yet, bad data were entered by Denton. This is very telling, as it strongly suggests that either Denton didn’t know there was a problem, or he did know and seemingly did not care…or he did know and he did care and he did not know how to address the problem, so he just carried on and entered problematic reports. During Denton’s second and third winters as observer at Greenland Ranch, the minimum temperature reports were consistently much too warm. It seems impossible, or implausible, that the observer could not know that there were major problems with his temperature reports from about October, 1913 to March 1914; and again from October, 1914 to March 1915. These problems begin only a few months following the record maximums from July, 1913.
The Greenland Ranch temperature record continued to be problematic at times post-Denton. These issues will be updated later.
——
The next part will investigate the relationship between summertime maximum temperatures between Greenland Ranch (and other Death Valley stations) with those at surrounding desert stations. It will be shown that, during the hottest parts of the summer, that maximums throughout the region are closely related to elevation. The well-mixed lower and mid-levels of the troposphere allow lapse rates at or very near the dry adiabatic rate from the surface to above the altitudes of the highest mountains in the region. Thus, maximum temperatures change at a rate of about 5 degrees (F)/1000 feet. The hottest temperatures in Death Valley closely associated with the hottest temperatures at the closest stations in the region.
Feel free to comment or to send any questions you may have my way ! E-mail
bill@stormbruiser.com
Leave a Reply
You must be logged in to post a comment.