Investigating crop performance on urban green roofs using hyperspectral data

Abstract

This study explores the dynamic relationship between crop plant performance and variety of treatment effects (water levels, diversity, and Sedum) on extensive urban green roofs. Green roofs have gained recognition as a sustainable urban design strategy, offering benefits such as stormwater management, energy efficient, and improved aesthetic enhancements. However, environmental stressors can challenge plant growth and performance on these emerging extensive green roofs. To help cope with these stressors, certain planting techniques have been implemented on green roofs such as plants grown with Sedum and increasing diversity types. However, monitoring the effects of these treatments in a rapid, timely manner are still in its infancy. To monitor plant health (specifically bush beans (Phaseolus vulgaris)) on green roofs, this study utilizes hyperspectral remote sensing technology using two crucial optical parameters: red edge point (REP) and red well point (RWP) to offer new valuable insights into monitoring plant health on green roofs. The results found that both REP and RWP information can detect significant differences between the treatment effects of water treatment (high and low) (p = 0.006, p = 0.001) and plants with or without existence of Sedum (Crassulaceae) (p = 0.01, p = 0.07) but could not find significant difference in diversity effects (single crop or multiple crops) (p = 0.20, p = 0.45). The results of REP and RWP was also able to find that plants with Sedum under low watering treatments showed worse plant performance than plants without Sedum under low watering treatments (p = 0.006, p = 0.006). This provides insights that remote sensing technology can detect that there may have been a competition of resources under low watering treatments and that that REP and RWP has the potential for a future sophisticated green roof monitoring platform.