NYT Failing to Explain California's Drought Accurately
At the New York Times, Carrie Halperin and Sean Patrick Farrell have a video from July 5 of this year on "California's Drought, Explained."
The problem is that the drought isn't accurately explained. At about 40 seconds into the video, the narrator states that "so far this year, California has had only 20 percent of its normal rainfall and the state's snowpack, another crucial water source, is only 18 percent."
Who knows what the NYT is trying to communicate, but it is surely misleading its readers. While the statewide drought in California is severe, the NYT is most definitely overstating its severity. For all this paper's smugness on science communication skills, it often does a terrible job at conveying the facts without bias or error.
California's statewide precipitation to date this year (May is the latest month available) has been 9.32 inches, or 67% of normal -- which is 13.83 inches. That is a long way from somewhere around 20 percent that the NYT is suggesting. The 2014 precipitation to date is only the 24th lowest on record since 1895.
Seems like someone at the NYT did some wonky math. Regardless, there is no way to explain the NYT's data on California's precipitation this year.
And in case you are wondering, there has been no significant trend in California's statewide January to May precipitation total since records began 120 years ago. California also defines a "water year" that runs from Oct. 1 through Sep. 30. There are no significant trends since 1895 in either the California water year precipitation to date (until the end of May) or for the entire water year (see figure below; latest full year available is 2012-2013).
Consequently, any notions that climate change is leading to drier water years in California have absolutely no statistical support. But that doesn't stop the activists. Back in November 2013, Peter Gleick of the Pacific Institute published a graph of California water year precipitation from 1896 to 2013 (i.e., the same data I just showed above), included a trendline, and then claimed the following:
"As bad as that is, however, there is an even more worrisome trend. Go back to Figure 1 and look at the trendline. This line shows the trend in statewide precipitation from 1896 (when good records began), and it shows a decline. California precipitation has, on average, been declining, from an average of around 23 inches per year to around 21 inches per year -- a nearly 10 percent decline in the past 117 years. This could be the result of natural variability, disparities in measurements over time, or climate change -- we don't know for sure. But it is bad news in a region where growing populations and demands for water are already an economic, political, and social challenge."
This is nonsense. Both parametric and non-parametric tests for trends in California's water year precipitation since 1896 yield massively non-significant results (p=0.88 for the traditional parametric linear regression method -- almost a perfect non-correlation).
We've been down this substandard science road before most recently with Seth Borenstein at the Associated Press, and his use of non-statistically significant trends in an attempt to make it look like far more regions of the United States are warming than really are. It is as simple as this: if your trend test yields a statistically non-significant slope (i.e., p-value greater than 0.01 or 0.05, whatever your reasonable criteria for significance is), then you cannot reject the null hypothesis that there is no trend in the data with any level of confidence.
Take-home messages: (1) California is dry, but not as dry as the NYT suggests, and (2) there is absolutely no evidence of declining statewide precipitation trends in California since 1895 either on annual or water year bases.