On Irrigation and the Hydroclimate

The interface between climate modeling and water resource management has long interested me and has recently returned to the news (well, at least the science-related news) in the form of irrigation-induced precipitation. In a recent paper, Min-Hui Lo and Jay Famiglietti demonstrated that irrigation in the central valley of California may have significant second order impacts in the form of increased precipitation over the Colorado River basin. The described hydrologic feedback loop is entirely anthropogenic in that the same water originating in irrigation induced precipitation (along with much more water) is diverted from the Colorado River to Southern California (although not directly to the central valley).

(Lo and Famiglietti, 2012)

(Lo and Famiglietti, 2012)

This sort of anthropogenic, hydrologic loop may not be unique. Quite simply, it has been difficult to model such processes using existing models due to the lack of irrigation scheme in large-scale models.

Despite the number of studies investigating the impacts of irrigation on climate, irrigation is still often missing from major land surface models (LSMs). Researchers have circumvented this difficulty by manually adding irrigation to model soil moisture in a number of ways – prescribing current rates of irrigation to continue into the future, assuming saturation at all times, taking a threshold approach – and while these solutions begin to address the problem, they are still unable to capture potential nonlinearities or anthropogenic feedback cycles.

Identification of the equilibrium feedback loops (“equilibrium” being used somewhat loosely) is only the beginning of understanding how large-scale irrigation affects the hydroclimate. In fact, humans play a dynamic role in the hydroclimate that will vary both with climatic and economic changes. We provide a forcing the climate system in the form of application of soil moisture, depletion of ground water, depletion of surface water, creation of reservoirs and (as a result) land cover changes. These various components have often been investigated as independent phenomena, but in reality they are not only correlated but causal. As temperatures rise and potential evapotranspiration increases, reservoirs may become habitually depleted, surface water will become less available, ground water withdraws will become greater and, if crops cannot be sufficiently irrigated due to cost or water availability, land cover will significantly change leading to a number of feedback cycles.

As climate models become increasingly nuanced, one area that deserves adequate attention is the irrigation schemes of LSMs as they may apply both to the hydrologic cycle and to the land cover scheme in the LSM. Irrigation is particularly relevant not only because it has had demonstrated effects on regional climate, but it is an entirely anthropogenic climate forcing. The expansive nature of irrigation, which enables it to act as a sizeable climate forcing, is unlikely to change in the near future. However, the way in which irrigation is practiced may change.

There seems also to be value in reminding ourselves that we have affected our environment in such fundamental, tangible ways. While the idea that water management is a means of providing resources to a population in a hostile environment is a position oft held, I believe it is founded on a false dichotomy. Drawing the distinction between “natural” and “anthropogenic” forces acting on the hydrologic cycle is no doubt a useful one, but the tendency to treat these actors as having distinct spheres of influence to be studied in isolation is detrimental to the advancement of water management. Understanding how our constructed hydrologic environment interfaces with the natural hydroclimate is crucial to developing better informed water management policies. Climate scientists, engineers and risk managers need to continue making a concerted effort to collaborate at local, state and federal levels.