Quantifying Evapotranspiration from a Green Roof Analytically

Green roofs are a common stormwater control measure used to help mitigate increased stormwater runoff caused by urbanization. Green roofs can retain upwards of 90% of rain given the appropriate conditions. Green roof design parameters, such as substrate composition and depth and plant type, affect the retention capabilities of green roofs by storing and allowing greater extraction of water by plants for evapotranspiration (ET). Evapotranspiration has been studied extensively in agriculture to be able to estimate irrigation demands, but ET in stormwater management infrastructure is poorly understood. Measurements from a weighing lysimeter consisting of a 10.2 cm deep, 45.7cm x 45.7 cm green roof replica on 3 Sentran load cells, was compared with results from the standardized reference Penman-Monteith equation to obtain crop coefficients (kc) for future predictions of ET from green roofs. The years 2009 and 2010 contained very different weather patterns with 2009 being wet and mild and 2010 being dry and hot. The weight of the lysimeter was used to estimate the soil moisture content and determine the moisture stress (with stress coefficient, ks). Ascertaining the level of stress quantifies the reduction expected in ET due to water shortage and proved to be critical to obtaining reasonable crop coefficient values. The average stress in 2009 was 1.0 with monthly crop coefficients ranging from 1.0 to 1.4; in 2010, the average stress coefficients were between 0.6 and 0.7 and dropped as low as 0.2 with estimated monthly crop coefficients ranging from 1.5 and 1.7. The high water stress levels present in 2010 made it difficult to separate kc from ks, but use of the 2009 crop coefficients in 2010 with estimated stress levels did not produce viable results, indicating that ks may not be sufficient in representing severe stress conditions. Although the model ETc,adj in 2010 still compared well against the measured ET, ETlys, it is believed that better results could be obtained using a dual-crop coefficient approach, where transpiration and evaporation are modeled individually, and results moer similar to those from 2009 would be obtained. The work presented is a preliminary attempt at establishing kc values for use in future design work, and while still in need of validation, could prove useful for rough estimates of ET.