FINE-SCALE MODELING OF URBAN HYDROMETEOROLOGY IMPLEMENTING FULL DYNAMICS OF ATMOSPHERE-LAND SURFACE-SUBSURFACE PROCESSES
Links to Files
Permanent Link
Author/Creator
Author/Creator ORCID
Date
Type of Work
Department
Chemical, Biochemical & Environmental Engineering
Program
Engineering, Civil and Environmental
Citation of Original Publication
Rights
Access limited to the UMBC community. Item may possibly be obtained via Interlibrary Loan through a local library, pending author/copyright holder's permission.
This item may be protected under Title 17 of the U.S. Copyright Law. It is made available by UMBC for non-commercial research and education. For permission to publish or reproduce, please see http://aok.lib.umbc.edu/specoll/repro.php or contact Special Collections at speccoll(at)umbc.edu
This item may be protected under Title 17 of the U.S. Copyright Law. It is made available by UMBC for non-commercial research and education. For permission to publish or reproduce, please see http://aok.lib.umbc.edu/specoll/repro.php or contact Special Collections at speccoll(at)umbc.edu
Abstract
This work built and evaluated a fully-coupled urban atmosphere-land surface-subsurface model, WRF-PUCM-PF, by coupling Weather Research and Forecasting (WRF), Princeton Urban Canopy Model (PUCM), and ParFlow. The WRF-PUCM-PF model was compared to WRF-PUCM, in application to a small watershed (Dead Run, 10.8 km by 10.8 km) in Baltimore, Maryland (90-m grid resolution). WRF-PUCM-PF turned most of the generated rain over impervious surfaces into runoff and had a 2% increase in area-averaged soil moisture (SM), whereas WRF-PUCM gained 10 times higher SM through infiltration over highly urbanized areas. WRF-PUCM-PF's SM distribution was influenced by topography and land cover, whereas WRF-PUCM's SM distribution was similar to the accumulated rain distribution because lateral flow is neglected in the model. Next, the sensitivity of WRF's fine-scale urban simulation to (1) leading time in SM spinup in WRF (14, 7, and 4 days before the analysis period), and (2) the conversion algorithm of the National Land Cover Dataset (NLCD) Developed categories to WRF's urban categories, was evaluated for the Baltimore metropolitan area (97.2 km by 97.2 km), gridded at 150 m. Overall, starting 7 days earlier resulted in better performance in LST prediction by perturbing SM distribution and not diverging from atmospheric observations. WRF's conversion algorithm of NLCD had the strongest impact on WRF's LST prediction. Finally, the role of incorporating a realistically simulated SM distribution (50-m grid resolution) by ParFlow into WRF was evaluated and validated for Baltimore City (36 km by 36 km). ParFlow-simulated SM was injected into WRF and compared to a simulation using interpolated SM from the WRF parent (outer) domain output. Validated against Landsat 8 land surface temperature (LST) on August 22, 2017, 12:00 EDT, the model with ParFlow SM input performed significantly better by achieving smaller biases in area-averaged LST (0.41 ºC and 1.06 ºC) over non-urban and urban areas, compared to the model with the interpolated SM distribution (1.2 ºC and 3.2 ºC). All simulations in this study were in large-eddy simulation (LES) mode. Overall, this work documents the importance of incorporating realistic terrestrial hydrology dynamics in the simulation of urban SM distribution and microclimatic variation.