Basin-Scale Climate Change Analysis Using High Performance Computing and the Integrated Hydrologic Modeling Platform ParFlow-CLM

The parallel, scalable integrated hydrologic platform built around ParFlow simulates the terrestrial hydrologic system using a fully implicit solution of variably-saturated subsurface flow and overland flow. Land surface processes and dynamic subsurface-land surface interactions are captured through a coupling of a modified version of the Common Land Model (CLM) over the top layers of a ParFlow model. Designed for parallel implementation, ParFlow-CLM performs well in a high-performance computing environment and allows high-resolution simulation of regional hydrologic systems. In this study we demonstrate the use of the a coupled model of the San Joaquin River basin system (59,400 km²) as means to understand the impacts of rising temperatures on subsurface and land surface hydrology in the context of precipitation variability leading in to the recent drought.

As with many catchments in the mountain west, water resources in the San Joaquin River basin are vulnerable to a warming climate through diminished snowpack and increased evaporative demand. The impact of warming is not necessarily uniform over a complex landscape, such as that formed by the Sierra Nevada Mountains and the Central Valley, nor is it constant across a typical range of wet and dry years. To address this spatial and temporal variability, we simulate the terrestrial hydrology of a 5-year wet-dry cycle in the San Joaquin River basin under baseline (no warming), uniform 2°C warming, and uniform 4°C warming scenarios using the ParFlow-CLM integrated hydrologic platform. The precipitation timing, amount, and spatial patterns are identical across each scenario. At the basin scale, incremental warming partitions more precipitation into evapotranspiration (ET) and away from runoff, although the larger change in partitioning occurs between the baseline and 2°C warming compared to the 2°C and 4°C warming.  In general, the year-to-year variation in precipitation tends to be the dominant first-order control on runoff while ET is sensitive to both precipitation and temperature perturbations. The shift in precipitation partitioning under warming scenarios suggests the recent drought may be a good analogy for an average future climate year. Specifically, impacts from the 4°C warming baseline year simulation reduce streamflow volume to a level associated with the current-climate dry years. Additional analysis of spatial and temporal properties of changes in runoff, ET, and subsurface storage suggest this aggregate pattern is the consequence of a complex combination of behaviors and feedbacks within the system.