Evapotranspiração de Sedum rupestre em telhados verdes extensivos

Abstract

Green roof is a technique embedded in the design of Low Impact Development with recognized in the quantitative stormwater control in urbanized areas. The green roofs storage capacity is affected by several factors, being evapotranspiration one of the most important. Due to the lack of information related to green roofs evapotranspiration or even the role of this component in the water balance in these systems, this research was developed aiming to evaluate evapotranspiration rates of experimental extensive green roofs, vegetated with Sedum rupestre. The experiments were performed in subtropical Brazilian climatic conditions, and the observed values were compared to those resulting from typical predictive equations of evapotranspiration. The experiment was composed by six modular vegetated green roofs and four non-vegetated, installed and monitored at the Federal University of Santa Maria from July 2016 to February 2017. As a result, the median evapotranspiration was 2.6 mm day-1, with values ranging from 0.2 mm day-1 to 5.1 mm day-1; the median evaporation of the study period was 1.9 mm day-1, with values ranging from 0.1 mm day-1 to 9.5 mm day-1. Higher evapotranspiration rates were observed in summer (3.0 mm day-1), followed by winter (2.3 mm day-1) and spring (1.8 mm day-1), while evaporation, these values were 4.3 mm day-1 (summer), 1.9 mm day-1 (winter) and 1.0 mm day-1 (spring). In percentage terms, the highest total evapotranspiration loss (in relation to rainfall) were observed in July (49.0%), September (87.1%) and December 2016 (92.9%). Considering seasonal analysis, 49.8% of the winter returned to atmosphere as evapotranspiration, while in the spring and summer these values were 25.0% and 29.9%, respectively. Vegetated and non-vegetated modules had similar performance regarding flow control, with runoff coefficient of 0.47 and 0.43, respectively. The results of the research show that the average crop coefficient (Kc) and water stress coefficient (Ks) was 0.9, with an average monthly variation of 0.5-1.4 and 0.7-1.1, respectively, highly dependent on the stage of vegetation development and environmental conditions. In order to evaluate the ability of predicting evapotranspiration using classical equations, eight models were tested. Equations of Penman-Monteith-FAO56, Penman, Jensen-Haise and Turc equations presented reasonable daily results without including Kc and Ks to the reference evapotranspiration values. However, good adjustments were only observed in some periods. By adding the Kc and Ks coefficient to the predicted evapotranspiration values, the adjustments became substantially better, and the models of Penman, Slatyer-MclIroy and Priestley-Taylor presented excellent results for all periods analyzed. Regression models basead on climatological variables observed during the study were also development, which showed good adjustments only in few periods, and the relative humidity of the air stood out as an important explanatory factor of evapotranspiration. This work presents important contributions on the behavior of green roof´s evapotranspiration and serves as a stimulus for the development of complementary surveys, aiming at a better understanding of the technique and, consequently, its diffusion, especially in Brazil.