Addressing Carbon-Plant Relations in Climate-Scale Analyses
There is more to capturing all the dynamics necessary to fully reflect drought or drying/wetting trends at climate timescales than simply turning to a fully physical E0 measure: at these timescales, plant-carbon (P-C) relations will change, with significant hydrological consequences. The heart of the issue is that using current fully physical E0 formulations, including either of the widely used Penman-Monteith equations for E0 (Ep or ET0), implicitly fixes P-C relations: such formulations lock in the response of vegetation to the current climate for which the wind function (in the case of Ep) and the stomatal conductance (in the case of ET0) are calibrated. Doing so ignores the effects of climate-scale physiological changes of vegetation to increased CO2: that increased water-use efficiency (WUE) leads to a similar carbon uptake and photosynthesis for lower transpiration losses from the plant. A stark warning on using E0 formulations with fixed P-C relations in climate-scale analyses is sounded by Roderick et al. (2015), who survey the contradictory claims and assumptions of past analyses.
Exemplifying the danger of assuming fixed P-C relations, Feng and Fu (2013) found that the Penman-Monteith ET0 resulted in dramatic (230 mm yr-1) increases in ET0 under coupled model intercomparison project (CMIP) phase 3/RCP8.5 drivers (from mean drivers from 27 models) in the period 2070-2099 relative to 1970-1999. Combined with a low Prcp increase (41 mm yr-1), this led to decreases in the aridity index (Prcp/E0) over much of the terrestrial land surface and an increase in aridity at a global scale. Rising T was shown to have led to an ET0 increase via an increase in vapor pressure deficit of 7-9 percent/K over land with the other drivers remaining inconsequential. However, Roderick et al. (2015) consider that this type of analysis lies at the root of the "warming is drying" message previously discussed. In contrast, they examined the aridity of the land surface at climate scales using the aridity index, noting that, despite regional increases and decreases, there has been little overall change in global aridity since 1948. They contest the commonly held notion that a warmer climate leads to a more arid land surface, and claim to have resolved the mismatch between observations and modeling in such a way as to align with the geological record. In doing so, their study highlights an example of a hidden modeling problem in at least the CMIP3 GCMs discussed: that of a fixed stomatal conductance.
Instead, Roderick et al. (2015) suggest assessing agricultural and ecological drought or aridity trends at climate scales using a new approach drawn from scientific communities in agricultural, ecology, and forestry. They propose using vegetation data from GCMs that do not rely on a fixed stomatal conductance parameter or wind function for estimating E0. Instead, the aridity index is based on the ratio of gross primary productivity (GPP) to WUE. In doing so, they note that CMIP3 and CMIP5 models show, on global average, a warmer climate being less arid for both meteorological (low Prcp) and hydrological (low Runoff) drought, and they suggest, that as GPP increases with increasing atmospheric CO2, global agro/ecological aridity will decrease. Their results resolve GCM modeling with remotely sensed observations and the geological record (i.e., they resolve the global aridity paradox). But questions remain (see Section 11.5).