A novel use of remote sensing to identify favorable populus deltoides recruitment areas at the reach-scale in a humid subtopical system

Abstract

The degradation of riparian systems due to the influence of human activity has had a significant impact on biotic assemblages commonly found in these unique zones around the world. The regulation of waterways via damming and withdrawal is one example of how human activity has altered river hydrology leading to adverse ecological impacts on riparian structure and function. Riparian structure and function are constrained by many factors including, among others, a natural flow regime to sustain biotic assemblages throughout a river’s reach. A natural flow regime includes seasonal variations in water inputs as well as changes brought about by disturbance events, such as floods. Therefore, my study will aim to highlight a high water event and the subsequent water recession directly related to the recruitment of many early successional species commonly found in riparian areas, specifically the Populus species. The study area will focus on a 24-km regulated segment of the Red River, bordering Oklahoma and Texas. Given my area of interest, only the eastern cottonwood (Populus deltoides) will be considered in the analysis. A novel methodology relying on remotely sensed data, specifically synthetic aperture radar data, is presented here. My methodology was able to isolate the water surface extents from the chosen high water event while quantifying the corresponding recession rates over a period of 36 days. A spatially explicit map corresponding to the favorability of Populus deltoides recruitment was produced according to a strict interpretation of established thresholds. I subdivided the 24-km reach of the Red River into 23 1-km segments to increase computational efficiency whereby 5 segments resulted in being classified as a marginal cottonwood recruitment areas and 1 segment classified as being favorable for cottonwood recruitment. My results from this study suggest the method performed very well given the chosen disturbance event was marginal at best for cottonwood recruitment. Continued refinement in the methodology to improve performance as well as addressing uncertainty in the data will need to be explored further to produce a model appropriate for cottonwood recruitment.