This project investigates the relationship between the dominant spatial scale (in the scale range 5-100 km) of heterogeneity in surface energy fluxes and the characteristic spatial scale of the diurnal, lower-tropospheric mesoscale circulations forced by this heterogeneity. Quantifying the links between these two scales is important for improving our basic understanding of coupled land-atmosphere dynamics and thermodynamics, for choosing the appropriate resolution in numerical mesoscale model experiments and mesoscale observational network design, and for developing parameterizations of subgrid-scale boundary layer and convective processes for large-scale models.

Past work of this kind has considered idealized surfaces with simplified distributions of surface properties. This investigation considers real (observationally based) surfaces from two distinct regions with extensive surface spatial heterogeneity in this scale range: Oklahoma/Kansas in the U.S. Central Plains and Rondonia, Brazil in the Amazon basin. We note also that a significant fraction of the mesoscale landscape heterogeneity in both these regions is the result of human activity (crop cultivation and deforestation, respectively). We carry out high-resolution simulations with the Regional Atmospheric Modeling System (RAMS) over these two regions of case study days from the GCIP Enhanced Seasonal Observing Period (ESOP) of 1995 (specifically, July 1995) and the Rondonian Boundary Layer Experiment (RBLE) of August 1994.

Our major findings include the following. While there are a variety of spatial scales that make a significant contribution to the overall surface flux variability, the peak atmospheric response occurs at a distinct, intermediate scale. For example, while the GCIP cases have maximum surface variability at relatively large scales, and the RBLE cases have maximum surface variability at relatively small scales, the peak boundary layer dynamical response occurs at intermediate horizontal scales that are similar for both regions: i.e., there is a preferred scale. This scale selection arises out of a competition between two dynamical factors. Surface sensible heat flux patches that produce mesoscale circulations able to grow into the afternoon and dominate the local dynamics are: (i) large enough, and have a great enough flux differential with their surroundings, to maintain strong horizontal pressure gradients in the face of mixing and advection; (ii) small enough so that opposing flows into the center of the patch converge quickly, thereby allowing the associated vertical motions to intensify earlier and out-compete other, slower-developing circulations. The mean surface sensible heat flux is also an important factor, as it helps determine whether the nascent mesoscale circulations will intensify quickly enough so that their development coincides with the overall maximum rate of buoyancy creation and the time of maximum boundary layer instability. This preferred scale appears to be a relatively weaker function of large-scale background meteorology compared to the characteristics of the surface sensible heat flux distribution. Methods for predicting the atmospheric scale and dynamical response from the pattern of surface fluxes are also discussed.