NSF Project: Data Complexity and Spatial Scaling
Prediction Accuracy and Implications for Emerging Landscape Paradigms
Impact of data loss on downscaling accuracy
The integration and examination of data collected at different scales necessitates aggregation, which reduces data quality and introduces statistical bias. These biases, collectively known as the Modifiable Areal Unit Problem (MAUP), result from different forms of data loss (e.g., compositional/configurational heterogeneity) that occur during aggregation. Building on recent theoretical and methodological advancements in spatial science, remote sensing, and landscape ecology, this research investigates the impact of heterogeneity and data loss on scaling across geographic contexts.
Patch vs Surface Paradigm
Conceptualizing and analyzing landscapes as mosaics of discrete patches has formed the fundamental structural basis of landscape ecology for decades, but landscape ecologists have begun recognizing that the patch-mosaic model may not be optimal for representing continuous spatial heterogeneity. The objective of this research is to create foundational linkages between the patch-based and surface paradigms to aid adoption of gradient surface analyses.
Forest fragmentation: a fuzzy perspective for a clear cut problem
Forest fragmentation is an on-going threat, especially in the Eastern U.S. where the prevailing pattern of dispersed, low intensity development penetrates intact forest, increasing the amount of wildland urban interface (WUI). Many methods to measure forest fragmentation rely on thematic land cover maps, which ignore the gradient nature of land covers. Fuzzy methods offer an alternative for modeling land cover as a gradient, but maps produced by these methods are not compatible with typical tools for measuring fragmentation. This study aims to bridge gaps between traditional fragmentation metrics and gradient datasets to better monitor forest fragmentation.
Urban scaling
Natural systems aren't the only geographical phenomena that exhibit scaling relationships and laws. Certain characteristics of cities and city systems have also been found to scale according to power-law functions. PhD student Gustavo Ovando Montejo is investigating whether certain measures of landscape configuration scale for a system of cities, and whether those relationships are universal across different countries and modes of urban development.