Rainfall is the most influential factor affecting groundwater and surface water regimes. Spatial and temporal characteristics of rainfall events significantly impact the partitioning of rainfall into runoff, infiltration, and groundwater recharge, and thus the dynamic water exchange between groundwater and surface water. Integrated hydrologic models simulate surface water and groundwater processes and their dynamic interaction using rainfall directly as an input. A network of rain gauges coupled with a spatial distribution process is often used to define the temporal and spatial distribution of rainfall over a model domain. The ability for a rain gauge network to capture key characteristics depends on the density and recording frequency of the network and the dominance of convective storms. Where the surface and groundwater systems are dynamically coupled, an integrated model is a valuable tool to evaluate changes in groundwater and surface water resources due to many causes including rainfall variability and changing climate.
Using a 4,000 square mile region of west-central Florida as a case study, integrating rain gauge data with NEXRAD radar rainfall data is demonstrated to enhance the estimation of historical rainfall and has resulted in improved hydrologic responses from an integrated hydrologic model of the region. The model application is based on the Integrated Hydrologic Model code which dynamically couples the HSPF surface water model and the MODFLOW groundwater model codes. The integrated model was previously calibrated using historical rainfall inputs that were estimated from rain gauge data only, with spatial distribution approximated by the Thiessen polygon method. Using a Bayesian statistical approach (described in a companion abstract), rain gauge and NEXRAD radar data sources were integrated to improve the estimation of historical rainfall from 1995 to the present time. Subsequently, integrated hydrologic model simulation was performed using the integrated rainfall inputs. The results show improved matching with observed flows and levels data that were used in model calibration. Model parameters were adjusted to further improve model performance. Model performance using the gauge-only and NEXRAD-only rainfall estimates is compared to the performance of applying the improved rainfall estimates.