A major theme of this research is translating remotely sensed data into ecologically meaningful results. This researcher is currently engaged with various stakeholders to evaluate delineated areas via a suite of metrics particular to each project. One of the projects involves evaluating the performance of a portfolio of land acquisitions. To do so, the researchers are combining many inputs in order to develop indices that meet the needs of the stakeholder and reflect their values. Ultimately, they will be scoring each acquisition according to the metrics they develop, in order to report how well the acquistions met the goals of the organization. Another project requires modeling forestry metrics and classification of forest stands. The latest project involves the evaluation of protected areas in order to assess the need for land management.
The goal is maximal efficiency and wise use of resources to complete analyses. Supercomputing access is beneficial because the analytical scale is statewide. Often the analytical input data is comprised of multiband rasters with a spatial resolution of 40m2 or less. Relatedly, dozens of predictor variables can go into a model depending on the question. Usually, summarization needs to happen for of millions of land parcels, so often analysis involves dividing the state into manageable areas (e.g. counties) for processing. Due to the resulting complexity of the models and large spatial data involved, MSI resources are necessary.