A method was developed for extending a fine-scaled forest gap model to a watershed-scaled landscape, using the Eastern Cross Timbers ecoregion as a case study for the method. A topographic wetness index calculated from digital elevation data was used as a measure of hydrologic across the modeled landscape, and the gap model modified to have with a topographically-based hydrologic input parameter. The model was parameterized by terrain type units that were defined using combinations of USDA soil series and classes of the topographic wetness index. A number of issues regarding the sources, grid resolutions, and processing methods of the digital elevation data are addressed in this application of the topographic wetness index. Three different grid sizes, 5, 10, and 29 meter, from both LiDAR-derived and contour-derived elevation grids were used, and the grids were processed using both single-directional flow algorithm and bi-directional flow algorithm. The result of these different grids were compared and analyzed in context of their application in defining terrain types for the forest gap model. Refinements were made in the timescale of gap model’s weather model, converting it into a daily weather generator, in order to incorporate the effects of the new topographic/hydrologic input parameter. The precipitation model was converted to use a Markov model to initiate a sequence of wet and dry days for each month, and then daily precipitation amounts were determined using a gamma distribution. The output of the new precipitation model was analyzed and compared with a 100-year history of daily weather records at daily, monthly, and annual timescales. Model assumptions and requirements for biological parameters were thoroughly investigated and questioned. Often these biological parameters are based on little more than assumptions and intuition. An effort to base as many of the model’s biological parameters on measured data was made, including a new technique for estimating optimal volumetric growth rate by measuring tree rings. The gap model was set up to simulate various terrain types within the landscape.

Throughout its current range, the Cross Timbers forest ecosystem is vulnerable to land-use change. In this study, we examined the surrounding land use matrix on the vegetation structure, composition and regeneration of six Cross Timbers forest fragments in Denton County, Texas (north of the Dallas-Fort Worth metroplex). Two fragments adjacent to agricultural land, two to residential neighborhoods, and two formally protected forest sites were selected. In summer 2015, five 100 m2 plots were randomly established in each fragment at least 200 meters from the edge. In each plot, all live and dead trees ≥ 3 cm diameter were identified and their height and diameter at breast height (DBH at 1.3 m aboveground) measured. Evidence of dumping (presence of trash) was recorded as an index of human frequentation. Differences in vegetation structure among the forest fragments were found. Most notably, fragments adjacent to agriculture contained 25% to 50% fewer trees per hectare than all other sites (Kruskal-Wallis, p < 0.02), especially trees <10 cm DBH. However, residential fragments had fewer trees that were ≥15 cm DBH compared to the other fragments, indicating that these are the youngest of the forest patches surveyed. Trash was observed in 60% of plots surveyed at residential forest sites, showing high levels of human frequentation compared to the protected and agricultural forest sites. Agricultural sites contained the lowest number of recorded tree species and were most similar to each other, sharing 91% of species. These findings indicate that surrounding land use affects forest structure and composition, consequently affecting valuable ecosystem services including wildlife habitat, aesthetics and recreation.