A new paper highlighting the importance of the urban land surface representation in numerical weather prediction (NWP) models has been accepted by the Journal of Geophysical Research. For this study the authors coupled an explicit urban canopy energy balance model with a photosynthesis based land surface model. These land surface models were then in turn coupled to a high resolution NWP model. This modeling system was used to simulate a mesoscale convection and precipitation event that was observed in the vicinity of Oklahoma City during a field experiment. The authors conclude that considering urban land surface explicitly improved the ability of the model to simulate precipitation and other model features. Without the explicit consideration of the urban model, the coupled system underpredicted rainfall, and had errors in the the location of convection as well as in accurately simulating the intensity of the temperature differences due to the urban heat island. Current NWP models- including GCMs (and also, therefore, climate models) do not explicitly consider urban areas and could have large errors due to this neglect in and around regions which have urban concentrations.
Dev Niyogi, Teddy Holt, Sharon Zhong, Patrick C. Pyle, Jeffery Basara, Urban and Land Surface Effects on the 30 July 2003 MCS Event Observed in the Southern Great Plains, Journal of Geophysical Research, accepted.
The abstract reads,
” The urban canopy of excess heat, water vapor, and roughness can affect the evolution of weather systems, as can land-vegetative processes. High-resolution simulations were conducted using the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS(r) ) to investigate the impact of urban and land-vegetation processes on the prediction of the mesoscale convective system (MCS) observed on 30 July 2003 in the vicinity of Oklahoma City (OKC), Oklahoma. The control COAMPS model (hereafter CONTROL) used the Noah land surface model (LSM) initialized with the Eta Data Assimilation System and incorporates an Urban Canopy Parameterization (UCP). Experiments assessed the impact of land-vegetative processes by: (1) adding a canopy-resistance scheme including photosynthesis (GEM) to the Noah LSM, and (2) replacing the UCP with a simpler urban surface characterization of roughness, albedo, and moisture availability (NOUCP).
The three sets of simulations showed different behaviors for the storm event. The CONTROL simulation propagated two storm cells through the OKC urban region. The NOUCP also resulted in two cells, although the convective intensity was weaker. The GEM simulation produced one storm cell west of the downtown region, whose intensity and timing were closer to the observed. To understand the relative roles of the urban and vegetation interaction processes, a factor-separation experiment was performed. The urban model improved the ability to represent the MCS, and the enhanced representation of vegetation further improved the model performance. The enhanced performance may be attributed to better representation of the urban-rural heterogeneities and improved simulation of the moisture fluxes and upstream inflow boundaries.
This study, although focused on numerical weather prediction, provides another example of the important role of the land surface as a climate forcing.