Summary of Climate-Scale Analysis Issues
In a broad assessment of climate-scale drought analyses, Trenberth et al. (2014) highlight the following issues that may lead to conflicting conclusions: uncertainties in input forcing—particularly Prcp and Uz—and in long-term ET estimates, where the differences between global land and regional estimates may be significant; and the difficulties in capturing the role of natural variability. Further, decadal trend assessments are unreliable without base periods long enough to capture natural variability—particularly when accounting for the effects of ENSO, the Pacific Decadal Oscillation, and the Interdecadal Pacific Oscillation. While Trenberth et al (2014) state that simple, T-based E0 parameterizations may have merit (so long as their shortcomings are recognized), we feel that this may be an acceptable compromise position only for informed hydrological and agricultural science communities. Indeed, we have seen other communities—such as ecological science communities and water-resource managers and their decision-making stakeholders—drawing conclusions from simple off-the-shelf models without appreciating their data and modeling nuances. We recommend against the use of PDSI for long-term climate analyses and, further, that studies that use T-based E0 in long-term analyses should be ignored. Indeed, with the near-ubiquity of long-term data for all drivers of E0 there is little cause to use such metrics at almost any timescale.