Elizabeth Wesley graduated in May 2018 with a Master of Science in Geography and a Master of Urban Planning through the joint degree program offered at KU. She is continuing her studies through the department, pursuing a PhD in Geography. We asked her to share information on her master’s thesis, “Using Wavelets and Landscape Metrics to Understand the Effects of Greenspace Configuration on the Urban Heat Island in Kansas City.” Her advisor was Dr. Nate Brunsell.
Motivated by the desires to understand the relationship between form and function in urban areas and to bridge the gap between scientists and those responsible for designing the built environment, my master’s thesis investigated whether the pattern of urban greenspace affects the distribution of urban heat. The urban heat island (UHI) is a consequence of human modification of the environment where urban areas show increased temperatures relative to surrounding urban areas due to the lack of vegetation and the thermal properties of building materials. Studies show that increasing the amount of vegetation in an urban area can mitigate this effect but what I wanted to know was whether given a fixed amount of greenspace the pattern of vegetated patches mattered as well.
Measuring pattern is complicated by issues of scale. Pattern changes depending on the boundaries with which you define it and and changing extent can change the statistical relationships between variables. In order to investigate the relationship between greenspace pattern and the UHI in the Kansas City area at a biophysically relevant scale, Dr. Brunsell and I conducted a multi-resolution wavelet analysis of land surface temperature (LST) to determine the dominant length scales of LST production. We used these extents for the calculation of landscape metrics to quantify the pattern of greenspace.
We then built regression models with these metrics to investigate whether, controlling for the percent of vegetated cover within the area, patch size, fragmentation, shape, complexity, and/or proximity can help mitigate heat island effects in urban areas. We found that increased edge density and patch density both decrease LST and increased patch cohesion increases LST. This indicates that more and complex patches of greenspace and dispersed rather than clustered greenspace can effectively mitigate UHIs. We also found that the negative relationship often reported between patch size and LST is an artifact of the relationship between increased percent vegetated and LST. By using the dominant length scales of LST we not only demonstrate that aggregation and shape complexity are important configuration factors to consider in designing urban greenspace, we also provide a methodology for robust biophysically-based analysis of urban landscape pattern.
PHOTOS: Top: This plot illustrates the linear relationships between LST and each metric across all vegetated classes and within each vegetated class. The color is the R-squared value; if a tile is grey that means the relationship is not significant at the 0.05 level. The number is the slope of the relationship. Middle, Ms. Wesley presenting her research at the 2018 American Association of Geographers (AAG) Conference in New Orleans, Louisiana. Bottom: The dominant length scale of each pixel represents the spatial scale at which the LST at that pixel was produced. The upper images show LST and the bottom show the dominant length scales for those dates.