Guest Post by Bob Vislocky
From the department of “use garbage temperature data to show causation by correlation” comes this recent Washington Post article which cites a report of a 60-fold decrease in the number of insects in Puerto Rico’s El Yunque rainforest … all because of climate change! Of course, this is “more widespread than scientists realized” the article states!
According to the report the maximum temperature increased by 4*F over the 38-year period of study (1976-2013), and that was the lone cause cited for the decline of insects. Of course this immediately set off the BS detector, so further investigation was initiated. First, maximum temperature data (summarized by month & year) was obtained for the closest NWS station to the El Yunque rainforest (San Juan) and for the closest active NCDC coop station which happens to be located just outside the rainforest (Juncos). These data are available from the San Juan NWSFO web site:
While there are no active NWS or Coop stations reporting within the El Yunque rainforest, there is one NWS Coop station inside the rainforest (Pico del Este) that reported maximum temperatures through 2004 (29 of the 38 years in the study). These data were compiled by NCDC over the years and maintained by the Southeast Regional Climate Center at the following web site:
Maximum temperature data for the three stations were organized in a spreadsheet, uploaded to the following link, and graphed as Figures 1-3 below.
Results show the max temperature trend in Juncos for the 1976-2013 study period was +0.0164*F per year. This amounts to a whopping gain of +0.62*F after 38 years. Meanwhile the max temperature trend in San Juan for the same period actually declined by -0.024* F per year! Finally, the trend from an actual site within the El Yunque rainforest (Pico del Este) was significantly negative by -0.075* F per year from 1976 to 2004. Clearly, there does not seem to be any obvious indication that the maximum temperature in the rainforest actually increased by 4* F claimed by the report, at least when using trusted weather data from nearby reliable sources.
So what weather data was used in the study? According to the original PNAS journal article that the Washington Post cited, the maximum temperature data was obtained from two research stations in the El Yunque rainforest (El Verde and Bisley) currently maintained by the USFS and USGS, and displayed in Figure 4 below.
The problem is that the data from those two stations should never have ever been used in this research study. For example, Bisley only started reporting in 1994, which means that the temperature record only covered the second half of the 38-year time period used to study the insects. Unless the authors had temperature data going back to the start of their study period in 1976 how in the world can they claim that the insect decline from their first expedition in 1976-77 to the second one in 2012-13 was due to climate change?
The situation for the observing station in El Verde is outright comical. This station has had a checkered history beginning with its odd choice of thermometer location (on top of a concrete roof). Roughly 25% of the data between 1976 and 2013 is reported as missing with no values provided. From 1976-1978 and 1987-1989 the daily observations were not real, but rather long-term average values were substituted. Moreover, prior to 1992 many of the max temperature observations were extrapolated from surrounding locations. Additionally, from 1989-1992 the temperatures reported were often that of the current reading at observation time instead of the maximum temperature, so adjustments had to be applied to correct those values. To top all of that there were instrument issues cited in the early 1990s that prompted replacement of the thermometer in September of 1992. The amount of corrupt data was so extensive prior to 1992 that the caretakers of the data set specifically state that this data is suspect and not valuable for interpreting long-term trends as it would pollute the later data record after 1992. Here are the links to the historical El Verde max temperature spreadsheets and more importantly the description of the data where the issues regarding the data sets are exposed.
But wait, there’s more! After the instrument change in September 1992, recorded maximum temperatures increased substantially at the station! As a result, a correction factor was applied to the data beginning in September of 1992 to make it compatible to previous data record; however the exact nature of the correction was not documented. Unfortunately, the correction factor ceased to be applied starting in 1997, so who knows what impact this had on the temperature trend. Lastly, even after the instrument change in 1992, over 35% of the daily readings are still reported as missing.
Despite all the warts on the post 1992 data, the caretakers of the El Verde max temperature database provided a handy tool to plot the post-1992 data on a graph (see link below) as a complement to the spreadsheet data they also provide.
Interestingly, when the data is plotted (see figure 5) it reveals a cooling trend contrary to what the authors describe for El Verde!!
By now it should be obvious that the max temperature data used by the researchers in this study was total garbage. There is no way they can justify a 4* F temperature rise from 1976 to 2013 and claim that climate change is killing off the insects inside the rainforest. Moreover, for the researchers to use the temperature data blindly for El Verde and Bisley when more reliable weather data was available from nearby trusted sites, at least as a double-check, is nothing short of gross incompetence. This is especially true considering that their entire thesis that insects are affected by climate change rests entirely on the integrity of the temperature data. The authors should also understand that correlations do not equal causations. Rather than blame climate change, perhaps the researchers should have looked elsewhere, such as land development, deforestation, tourist encroachment, pollution and invasive species, as the following articles point out.