All-Weather Ground Surface Temperature Simulation

Abstract

Thermal pollution from urban runoff is considered to be a significant contributor to the degradation of coldwater ecosystems. Impervious surfaces (streets, parking lots and buildings) are characteristic of urban watersheds. A model for predicting temperature time series for dry and wet ground surfaces is described in this report. The model has been developed from basic principles. It is a portion of a larger project to develop a modeling tool to assess the impact of urban development on the temperature of coldwater streams. Heat transfer processes on impervious and pervious ground surfaces were investigated for both dry and wet weather periods. The principal goal of the effort was to formulate and test equations that quantify the heat fluxes across a ground surface before, during and after a rainfall event. These equations were combined with a numerical approximation of the 1-D unsteady heat diffusion equation to calculate temperature distributions in the ground beginning at the ground surface. Equations to predict the magnitude of the radiative, convective, conductive and evaporative heat fluxes at a dry or wet surface, using standard climate data as input, were developed. Plant canopies were included for surfaces covered by vegetation. The model can simulate the ground surface and subsurface temperatures continuously throughout a specified time period (e.g. a summer season) or for a single rainfall event. Ground temperatures have been successfully simulated for pavements, bare soil, short and tall grass, trees and two agricultural crops (corn and soybeans). The simulations were first run for different locations and different years as imposed by the availability of measured soil temperature and climate data. Data came from sites in Minnesota, Illinois and Vermont. To clarify the effect of different land uses on ground temperatures, the calibrated coefficients for each land use and the same soil coefficients were used to simulate surface temperatures for a single climate data set from St. Paul, MN (2004). Asphalt and concrete give the highest surface temperatures, as expected, while vegetated surfaces gave the lowest. Bare soil gives surface temperatures that lie between those for pavements and plant-covered surfaces. The soil temperature and moisture model appears to model surface temperatures of bare soil and pavement with RMSEs of 1 to 2°C, and surface temperatures of vegetation-covered surfaces with RMSEs of 1 to 3oC. The plant canopy model used in this study, based on the work of Best and Deardorff, provides an adequate approximation for the effect of vegetation on surface heat transfer, using only a few additional parameters compared to bare surfaces. While further simplifications of the model are possible, such simplifications do not reduce the number of required input parameters, and do not eliminate the need for estimating the seasonal variation of the vegetation density. A model for roof temperatures was also developed, based on the surface heat transfer formulations used for pavement. The model has been calibrated for both a commercial tar/gravel roof and a residential roof. Compared to pavement, the roof surface reach similarly high maximum temperatures, but reach lower minimum temperature at night cool due to their lower thermal mass.