A robust projection of all climate models is the tendency for wet regions to become wetter and dry regions to become drier as the climate warms under increased greenhouse gases. This presents a…

This is a $388,425.90 grant award from the National Aeronautics and Space Administration (NASA) to the University of Miami Rosenstiel School of Marine and Atmospheric Science (RSMAS) Division. The grant is for developing a new statistical downscaling method to improve regional precipitation projections from global climate models, and applying these improved forecasts to water management planning in South Florida.

The project aims to address challenges in using global climate model outputs for local water resource planning, by correcting biases in the models' rainfall climatologies using satellite observations from NASA's Tropical Rainfall Measuring Mission (TRMM) and Global Precipitation Measurement (GPM) missions. The improved downscaled precipitation projections will then be used to drive hydrological modeling of the South Florida Water Management District's water management system. This work is intended to provide water managers with more reliable climate information to support seasonal to decadal planning decisions. No set-aside designation was mentioned for this grant award.

A robust projection of all climate models is the tendency for wet regions to become wetter and dry regions to become drier as the climate warms under increased greenhouse gases. This presents a clear problem for water resource managers as future rainfall patterns may be well outside the historical range of variability on which many managed systems are based. Such is the case in south florida, where multi billion dollar efforts to restore the everglades ecosystem and upgrade municipal water management facilities are underway without explicit consideration of future rainfall change due to the absence of credible projections. Because of their large regional biases, rainfall simulations from coupled global climate models (gcms) have proven to be inadequate for direct use by water resource managers in this region. Moreover, gcms differ substantially in their prediction of regional changes in precipitation. Recent research has shown that much of the intermodel spread in regional projections of precipitation change is due to differences in the models climatological distribution of rain. The fundamental premise of this proposal is that correcting for spatial displacements in the gcms climatology of rainfall using satellite observations from trmm (and GPM) will yield more consistent and more accurate regional predictions of precipitation change. The veracity of this hypothesis will be tested in two ways: first by evaluating the reduction in spread in the downscaled model projections of future rainfall change, and second through their application for water management in south florida. Many gcms simulate recognizable rainfall patterns, however they are often displaced spatially due to regional biases in the models climatology (e.g., SST biases, etc.). Existing statistical downscaling techniques do not have an explicit method to account for spatial displacements in the base climate. Here we propose a new approach for statistically downscaling GCM predictions that goes beyond traditional quantile mapping by first performing a spatial remapping to explicitly adjust for regional displacements in their climatologies using trmm (and GPM) measurements in conjunction with a spatial feature mapping algorithm. This algorithm has previously shown to be effective in mapping pattern displacements in other geophysical fields. By statistically correcting for biases in the spatial structure of the models unperturbed precipitation climatology, we can improve the consistency of model projections of precipitation change at the regional scale and provide improved products for planners and stakeholders. To accomplish this goal, we will complete the following tasks: 1) generate high resolution regional simulations using NASA trmm and GPM measurements to provide target climatologies for statistically downscaling climate forecasts from the new gfdl seasonal forecast model gfdl flor and the CMIP5 ensemble of GCM simulations. 2) use the high resolution downscaling of precipitation change to force the south florida water management model (sfwmm) and evaluate the utility of the downscaled precipitation for seasonal to decadal planning and management options within this framework. A key element to this work is the collaboration with south florida water management district (sfwmd). The hydrological model they have developed for the south florida region takes into account the diverse stakeholders incorporates the management processes that satisfy policy based rules (both existing and proposed) to meet flood control, water supply and environmental needs. Management options will be tested within this model framework using downscaled projections of precipitation change on seasonal to decadal time scales. This work addresses the PMM objective to develop improved downscaling of satellite precipitation information for hydrological modeling and prediction.