Current Treatment of ET and E0 in Drought
There are fundamental conceptual problems with traditional and current practices with respect to ET and E0. First, we recognize that the treatment of ET can be more burdensome than that of E0. ET is a difficult quantity to retrieve from remote sensing (RS): generally, various RS-derived datasets (e.g., land surface temperature [LST] or vegetation information) and meteorological datasets (e.g., Uz and T) must be combined in a physically based framework to estimate ET, usually indirectly from methods based on the relationship between sensible heat flux (H) and LST. Additionally, ET is difficult to validate as observations are not readily available, and when they are (e.g., from eddy-covariance [EC] tower platforms), they are only representative of small areas. The comparison of ET observations with RS retrievals is further complicated by scaling considerations of the EC tower footprint and incomplete energy closure of the EC observations. It should be noted that for drought applications, some of these issues can be minimized by transformation of ET retrievals into anomaly space; however, the need for long-term time series of RS-derived ET to ascertain the background mean state can be a further complication.
These complications explain why many traditional and current droughtmonitoring operations estimate ET indirectly, using land surface models (LSMs) that constrain an estimate of E0 by some measure of 0 (such as soil moisture [SM]), or:
where 0 may commonly be parameterized in an LSM (as in the Sacramento soil moisture accounting model). Many current or legacy hydrological and drought products use simplified parameterizations of E0 that are often buried deep in LSMs and that neglect much of the variety of forcings shown in Equation 11.2, instead relying solely on T or on some combination of T, q, and/or incident solar radiation (Rd). At most, such treatments examine drought as a function of Prcp and T only. These cannot adequately represent the physics of ET and E0 or their interrelations in drought, and this has severe consequences for the estimation of short-term variability vital to analysis of drought and of secular and climate-scale trends. The most charitable reading of such treatments of E0 is that they still examine drought as an imbalance of supply (Prcp) and demand, where T is used as a proxy for E0 and hence for demand. This short-cutting treatment of the drivers of drought is the central issue with current state-of-the-art monitoring and outlooks (as exemplified in the monthly and seasonal drought outlooks from the NOAA Climate Prediction Center [CPC]).