Impacts of land-atmosphere interactions on regional convection and rainfall
In this dissertation, interactions between land-surface heterogeneities, land-atmosphere coupling, and moist convection and related mesoscale circulations were investigated in four major studies to improve and advance the understanding of high-resolution model simulations of regional convection and precipitation. A number of short-term (i.e., 24-48 hours) retrospective numerical experiments were conducted over a variety of land-atmosphere coupling hotspot regions across the globe. First, impacts of heterogeneous land surface on turbulent flow and mesoscale simulations were assessed. Experiments were conducted using the Weather Research and Forecasting (WRF) model coupled with a simple (slab) land surface model (LSM), a modestly complex Noah LSM, and a land data assimilation system (LDAS) with detailed surface fields. Three heterogeneity length scales: 1, 3, and 9 km, were employed to alter land cover and land use. The response of high-resolution model simulations’ to spatial scales changes of land-surface heterogeneity by modification of land-surface properties and changes in land-surface representation were investigated. Results indicate that both land-surface parameterizations and surface heterogeneity affect model simulations, and the impact of land-surface parameterizations is found to be more important, particularly for low frequency (f < 10−4 hz) eddies and mesoscale circulations. Replacing a simple slab land model with more detailed land surface models (LSMs) (e.g., Noah or High-Resolution Land Data Assimilation System) can help reduce uncertainties in the simulation of surface fluxes which may be greatly affected by land-surface heterogeneity via improved turbulent characteristics over heterogeneous landscapes. An important result that emerges from the analysis is that the impact of land-surface heterogeneity on atmospheric feedbacks can be detected in mesoscale circulations that are roughly four times the heterogeneity spatial scale. It follows that the heterogeneity length scale that can influence mesoscale circulations would be a function of grid spacing in the model. Second, the role of land-atmosphere coupling over regions with relatively strong coupling between land-surface conditions and moist convection were assessed. The need for adopting a dynamic coupling strength within the land surface model was assessed by analyzing rainfall events and impacts of land-atmosphere coupling using the Noah land model and WRF model simulations over the U.S. southern Great Plains (SGP), Europe, northern India, and West Africa. Land-atmosphere coupling strength impacts on model parameterizations (i.e., land surface processes, PBL dynamics, and moist convection) were quantified and the range of regional variation in the coupling coefficient for model simulations was documented. Results indicate that the adoption of a dynamic land-atmosphere coupling formulation helps improve the simulation of surface fluxes and the resulting atmospheric state, thus dynamic coupling shows promise in modulating model results and improving convective system simulation and precipitation forecasts. For the four regions, the surface coupling coefficient does not affect the general location but could improve the intensity of simulated precipitation. Results highlight that there is high uncertainty in land-atmosphere coupling and the results from this and prior studies need to be considered with caution. In particular, zones identified as coupling hotspots in climate studies and their coupling strength would likely change depending on the model formulations and coupling coefficient assigned. Third, impacts of an updated convection scheme on high-resolution precipitation forecasts were assessed. At high resolution spatial scales, precipitation biases and errors can occur due to uncertainties in initial meteorological conditions, grid-scale cloud microphysics schemes, and/or subgrid-scale convection schemes. To reduce precipitation biases and uncertainties, scale-aware parameterized cloud dynamics were introduced to high-resolution forecasts by making several changes to the Kain-Fritsch (KF) convection parameterization scheme (CPS) in the WRF model. These changes include subgrid-scale cloud radiation interactions, a convective adjustment timescale, the cloud updraft mass flux impacting grid-scale vertical velocity, and a LCL-based methodology for parameterizing entrainment. This updated KF (UKF) CPS allows the convection scheme to facilitate a smooth transition from parameterized cloud physics to resolved grid-scale cloud physics across different grid resolutions. Results indicate that (1) high-resolution precipitation forecasting is more sensitive to the source of initial conditions than to grid-scale microphysics or convective parameterizations, and (2) the UKF CPS greatly alleviates excessive precipitation at 9 km grid spacing and improves results at 3 km grid spacing as well. In the last part of this dissertation, impacts of land-atmosphere-convection interactions on regional precipitation intensity and variation in the WRF model were assessed. Sensitivity experiments including effects of LSM, land-atmosphere coupling strength, and CPS on the fields of precipitation, surface scalars, and convection reveal that including a more detailed land surface parameterization, a dynamical surface coupling strength coefficient, and UKF CPS together, improves mesoscale simulations of several meteorological and convection parameters in the short-term high-resolution WRF model, increasing accuracy about 40% for precipitation intensity forecasts. (Abstract shortened by UMI.)
Harbor, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our