**1. Introduction**

During the past decade, green infrastructure (GI) gradually becomes a favorable concept to be associated with sustainable solutions to manage firstly water then later energy and food nexus in the urban environment. Traditional drainage infrastructure (often referred to as gray infrastructure) makes use of pipelines to rapidly export stormwater out of urban domain and then mitigate the rising flood risk induced by the expansion of impervious surface through urbanization. This water deficit then has to be resolved by importing high-quality potable water back into cities for irrigation and other uses [1]. In contrast to gray infrastructures with dull appearance and often hidden under covers, the visible components and lively forms make GI a more persuasive concept that is easily accepted and appreciated by the public. As a bridge connecting the water and energy cycles, evapotranspiration (ET) affects the overall performance of GI and will only receive more attention in the near future when more sub-disciplines can be taken into consideration.

The term green infrastructure emerged in the United States in the 1990s representing a network of green space stitching together the fragmented urban areas [2]. Its function in the field of stormwater management was widely realized only until the last decade, but the scope of GI quickly expands to involve other urban drainage terms such as Low Impact Development (LID), Best Management Practice (BMP), Stormwater Control Measure (SCM), Water Sensitive Urban Design (WSUD), Sustainable Urban Drainage Systems (SUDS), and Alternative Technique (AT) or Technique Alternative (TA) [3]. Besides the vegetated formats like green roof, bioretention, and vertical greenery systems [4, 5], GI also evolves to include other nonvegetation-based devices such as permeable/porous pavement and rainwater harvesting system designed for places, where vegetated GI is impractical to use due to heavily polluted runoff or the competing drinkable water demand [1]. More broadly, conventional urban green space, e.g. urban lawns, forests, farmlands, parks, and public gardens, has been used as a type of GI [6–9], owing to their capacity to promote retention and ET, as so-called natural water retention measures [10]. Recently, lakes and surface waters (so-called blue space) have futher been regarded as GI for improve local groundwater recharge, cooling, water purification, dust control, and a esthetics in an urban environment [11–13].

Evaporation happens directly from the water surface and porous media like soil, gravel, or permeable pavement. Transpiration occurs through the stomata on leaves as a subprocess of plant respiration. As two quantities are difficult to separate during measurement and modeling, they are often counted and treated as a total as referred to ET. As a stormwater management strategy, GI harvests and retains stormwater in the urban landscape [14], and then reuses and drains the captured water partly by ET. Evapotranspiration process also draws heat from surface when converting liquid moisture into vapor. It, therefore, provides a mechanism to mitigate the urban heat island effect [1]. The proportion of ET within urban water and energy budgets usually rises with vegetation coverage [8]. But only taking a small fraction of the urban surface, GI can provide an order of magnitude larger ET compared to the evaporation contribution from impervious surface [15]. Being spatially distributed within the street canyons, GI imports evapotranspiring "cool spots" into the urban ecosystem.

Previous research has given extensive reviews of the overall benefits of GI and listed ET as a process that requires more studies [16–18]. A critical review centering on ET process in GI, however, is lacking for GI community up to date. Therefore, this work endeavors to summarize the current research progress of ET with regards to GI and the knowledge gaps that restrict the development of the disciplines. Based on a survey of 100+ relevant peer-reviewed journal articles and book chapters in the previous decade, three current research areas are identified, which include the ecosystem service, measurement, and simulation of ET process from GI.
