Component characterization and predictive modeling for green roof substrates optimized to adsorb P and improve runoff quality: A review

摘要

This review is a synthesis of the current knowledge regarding the effects of green roof substrate components and their retentive capacity for nutrients, particularly phosphorus (P). Substrates may behave as either sources or sinks of P depending on the components they are formulated from, and to date, the total P-adsorbing capacity of a substrate has not been quantified as the sum of the contributions of its components. Few direct links have been established among substrate components and their physicochemical characteristics that would affect P-retention. A survey of recent literature presented herein highlights the trends within individual component selection (clays and clay-like material, organics, conventional soil and sands, lightweight inorganics, and industrial wastes and synthetics) for those most common during substrate formulation internationally. Component selection will vary with respect to ease of sourcing component materials, cost of components, nutrient-retention capacity, and environmental sustainability. However, the number of distinct components considered for inclusion in green roof substrates continues to expand, as the desires of growers, material suppliers, researchers and industry stakeholders are incorporated into decision-making. Furthermore, current attempts to characterize the most often used substrate components are also presented whereby runoff quality is correlated to entire substrate performance. With the use of well-described characterization (constant capacitance model) and modeling techniques (the soil assemblage model), it is proposed that substrates optimized for P adsorption may be developed through careful selection of components with prior knowledge of their chemical properties, that may increase retention of P in plant-available forms, thereby reducing green roof fertilizer requirements and P losses in roof runoff. (C) 2017 Elsevier Ltd. All rights reserved.