Impact of irrigation frequency and substrate composition on green canopy cover of extensive and intensive green roofs in Manhattan, Kansas (USA)

Abstract

Supporting the vegetation while minimizing supplemental irrigation is one of the major goals of sustainable green roof maintenance. Within the tallgrass prairie of the Flint Hills ecoregion (Great Plains, USA), supplemental irrigation and vegetation cover dynamics are critical due to the hot summers, high evaporative demand, and multi-week dry spells. In addition to supplemental irrigation, substrate depth and substrate type or mix are also critical variables influencing vegetation performance. There is insufficient research on how these variables influence vegetative cover. This thesis evaluates the canopy cover performance of shallower (extensive: ~10 cm deep) and deeper (intensive: ~20 cm deep) green roofs with two substrate types (lighter and denser) managed under two types of minimal supplemental irrigation (78 total gallons versus 26 total gallons added), using a mature, in-situ experimental green roof planted with prairie vegetation at Kansas State University. Collecting green canopy cover data is a methodological challenge in green roof research. This study applied multiple methods of green canopy cover analysis (NDVI data from UAS images, green cover data from UAS and iPhone images, and visual observations) to explore the strengths and shortcomings. In addition, this study explored substrate temperature and moisture dynamics and their relationships with green canopy cover. The methods did not always agree, and the disagreement extended to which factors appeared most influential. In the green-cover analyses, substrate type and irrigation frequency were the dominant controls and substrate depth was secondary, whereas the NDVI analysis of all 72 plots identified substrate depth as a major factor for vegetation health. This divergence, which reflects both the larger plot count available for NDVI and the different property each method measures, shows that the choice of measurement method materially affects the conclusions drawn and points to clear directions for future monitoring.