Front. Environ. Sci. Frontiers in Environmental Science Front. Environ. Sci. 2296-665X Frontiers Media S.A. 977084 10.3389/fenvs.2022.977084 Environmental Science Original Research Evapotranspiration and land surface temperature of typical urban green spaces in a semi-humid region: Implications for green management Li et al. 10.3389/fenvs.2022.977084 Li Xinhao 1 † Li Yiran 2 † Di Suchuang 3 Niu Yong 1 * Zhang Chuanjie 4 * 1 College of Forestry, Shandong Agricultural University, Taian, China 2 School of Soil and Water Conservation, Beijing Forestry University, Beijing, China 3 Beijing Water Science and Technology Research Institute, Beijing, China 4 College of Water Conservancy and Civil Engineering, Shandong Agricultural University, Taian, China

Edited by: Jifeng Deng, Shenyang Agricultural University, China

Reviewed by: Jing Qin, China Institute of Water Resources and Hydropower Research, China

Wenyi Dong, Chinese Academy of Agricultural Sciences (CAAS), China

 *Correspondence: Yong Niu, niuyong1988@126.com; Chuanjie Zhang, Zhangchuanjie1983@126.com

These authors have contributed equally to this work and share first authorship

This article was submitted to Drylands, a section of the journal Frontiers in Environmental Science

 24 08 2022 2022 10 977084 24 06 2022 28 07 2022 Copyright © 2022 Li, Li, Di, Niu and Zhang. 2022 Li, Li, Di, Niu and Zhang

This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

Ecological deterioration and water scarcity motivate people to seek a balance between ecological reserves and water consumption in cities located in Semi-humid regions. This study is designed to develop a method for evaluating cooling water-saving characteristics of green space structures. Land surface temperature (T S ), vegetation quantity (VQ) and evapotranspiration (ET) of typical green vegetation structures in 50 plots with an average area of 10,000 m2 were studied. Parameters were obtained based on data from three temporal remote sensing images; the surface energy balance algorithm for land (SEBAL) model, single-channel algorithm, and normalized difference vegetation index (NDVI) were utilized in the calculation. The relationships between the T S , VQ, and ET of green land were explored using regression analysis. A comprehensive index (EWI) was proposed to evaluate the cooling and water-saving abilities of different green space structures. Based on assessment results, species were classified into three: good (EWI ≥ 0.795), common (0.795 > EWI ≥ 0.419), and weak (EWI < 0.419). The characteristics of 17 species or species compositions with good cooling and water-saving abilities were summarized as: 1) a mix of the arbor, shrub, and herb, and 2) complete cover of green space by shrub.

 urban green space vegetation structure remote sensing inversion heat island effect water-saving No. ZR2016DB12 Natural Science Foundation of Shandong Province10.13039/501100007129 Introduction

Water is the source of life; efficient water use has significant implications for people who live in regions with little precipitation. Approximately 53% of China’s territory is made up of arid and Semi-humid regions. Ranked 121th in the world, China has annual per capita water resources of only 2,200 m3, and with over 400 cities, there is a serious water deficiency especially in the north. Currently, deficits existing between the water demand and supply are becoming the principal bottleneck to sustainable development. To alleviate the pressure on water resources, the Chinese government has implemented several significant water transfer and water-saving projects (Cai, 2008). However, with rapid urbanization, people have to cope with the problems brought about by the deterioration of the ecological environment in cities, such as water pollution, hazy weather, and heat island effect etc. This has resulted in economic losses.

As an important part of the urban ecosystem, urban green spaces can provide vital services to humans (World Resources Institute, 2005), and make both the ecosystem and society sustainable (Katherine et al., 2010). Although urban green spaces might be smaller in size compared to other land uses, the value of its ecosystem services is likely to be disproportionately higher depending on their locations (Perino et al., 2014). Maximizing the ecological value of green spaces partly dependents on a reasonable mix of landscape plants (Katherine et al., 2010). In 2011, the green coverage rate reached 38.62% in the cities of China (The National Afforestation Committee, 2019), but there are still problems such as unreasonable green space structure and inefficient use of water. Take Beijing as an example; in 2009, the green space area was 617 km2, urban green coverage rate was 43.5%, and annual irrigation water requirement was 2.2 m3 × 108 m3, nearly 10% of the gross city water consumption (Beijing Municipal Bureau of Statistics, 2019). The great water requirement of green space irrigation further aggravates the pressure on water supply in the cities, especially in the arid and Semi-humid regions. It is critical to understand the potential tradeoffs between service functional values of ecology and city water usage. In this regard, it is necessary to conduct a study on water consumption and ecological functions of vegetation structures. We also hope that our work can bridge the gap between ecological environmental quality and water resources in the arid and Semi-humid regions.

Hydrology is regarded as an important technological path in the study of water utilization and vegetation management (Masoud et al., 2007; Cheng et al., 2009; Yan et al., 2010; Zhao et al., 2010). This method laid the foundation for understanding plant water consumption. However, only a few studies focused on different green space structures with monoculture species. Also, the hydrology method is difficult when comparing large quantities of objects because of the unbearable cost (Li et al., 2009). Analogously, existing studies on service function of vegetation are mostly single case studies that focused on carbon sequestration, dust-retention, and oxygen releasing abilities, etc. (Katri et al., 2011; Yen and Lee, 2011; Liu et al., 2013; Zhao et al., 2013). Thus, there is a lack of comparative analysis. Meanwhile, spatial gradient analysis was widely used in landscape ecology studies, but it was not suitable for a plot scale (Kong et al., 2005; Tohru et al., 2011). The vegetation evapotranspiration, leaf quantity, and temperature of 50 different green space communities in the Beijing urban area can reflect the regional ecological carrying capacity to a certain extent. Above all, there would be some disadvantages of the above research methodology in a comparative study of ecological functions and water consumption of various urban green space structures.

The capacity and efficiency of environmental information collection for the Earth’s surface have been greatly improved by the development of remote sensing (RS) technology. Many algorithms have been applied to estimate vegetation evapotranspiration, structure, and land surface temperature at multiple spatial scales (Liu et al., 2007). Historically, there are three methods applied in the estimation of evapotranspiration: statistical methods (Jackson et al., 1977), energy residual methods (Granger and Gray, 1989; Hobbins et al., 1999), and numerical models. Recently, many quantitative RS studies have been conducted with SEBAL (Surface Energy Balance Algorithm for Land), SEBS (Surface Energy Balance System), and S-SEBI (Simplified-Surface Energy Balance Index) models (SEBS model and S-SEBI model were generated based on SEBAL model) (Idso et al., 1975; Bastiaanssen et al., 1998a; Roerink et al., 2000; Su, 2002), and good results were obtained. Since 1960, several algorithms including single-channel, split-window, multi-view single-channel, and multi-channel and multi-angle algorithms, have been proposed for the calculation of surface temperature with different RS data. (Becker, 1987; Roerink et al., 2000; Dash et al., 2002; Su, 2002; Jiménez Muñoz and Sobrino, 2003; Qin et al., 2010) used different algorithms to estimate the temperature of underlying surfaces.

This study investigated the vegetation evapotranspiration, leaf quantity, and temperature of 50 different green space communities in the Beijing urban area. The study sought to explore a method for evaluating the cooling and water-saving abilities of typical green space structures using RS and GIS and to provide a reasonable focus on green space structure and information in the management of urban green space, and research results are of significance in improving the ecological carrying capacity of the Semi-humid Region.

 Materials and methods Study site

The study sites were located in Beijing, northern China (Figure 1). Beijing, an ancient city with over 1,000 years of history, is a rapidly developing city, with more than eight million urban residents. The city consists of 14 administrative districts and four counties. In downtown, there are approximately 40 main parks.

Study area (bold-red bordered line).

The main study area, located in the northwest of the city (116˚14′38.6″E–116˚24′29.34″E, 39˚57′7.36″N–40˚2′51.3″N), includes parts of Haidian, Chaoyang, Changping, Dongcheng, and Xicheng district, with a gross area of 147.5 km2; the annual total rainfall ranges from 544.7 to 575.6 mm (Figure 1). The green space is mainly located in the National Olympic Park, Old summer Palace Park, summer Palace Park, Bajia Park, Haidian Park, Beitucheng Park, etc. Buildings, roads, water, green land, and bare land are the main land cover types present. Monoculture and mix-species arbor forests, shrubbery, grass and multi vegetation structures of tree-shrub-grass typically make up the green vegetation types.

 Remote sensing data

Three temporal remote sensing Landsat-8 images, which were generated on 12 May 2013, 13 June 2013, 1 Sep 2013, respectively, were used. The standard Landsat eight data products provided by the USGS EROS Center (http://landsat.usgs.gov/index.php) consist of multispectral image data acquired by both the Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS). One image scene consists of nine spectral bands with a spatial resolution of 30 m for bands one to seven and 9. Thermal bands 10 and 11 are useful in providing more accurate surface temperatures and are collected at 100 m.

 Meteorological data

The meteorological data were collected on 12 May 2013, 13 June 2013, 1 Sep 2013, respectively, and the hourly meteorological data involved in this study were collected from the Beijing Weather Observatory website (http://cdc.bjmb.gov.cn/shuju.asp), and this included solar radiation, rainfall, wind speed, temperature, and relative humidity. The collected meteorological data were used to evaluate the cooling and water-saving abilities of typical green space structures.

 Sample plots

50 plots with different vegetation communities were selected in the study area; their areas ranged from 6,400 to 12,100 m2 (average size, 100 m × 100 m), and away from buildings and water. Vegetation structure can be classified into pure arbor type, arbor-grass type, arbor-shrub type, arbor-shrub-grass type, shrub-grass type, and grass type. The vegetation structure information was collected by field survey, and it included: dominant plant species, species coverage, and sand density (as shown in Table 1).

Sample plot status.

Serial number Arbor layer Shrub layer Herb layer
Dominant species Coverage (%) Sand density Dominant species Coverage (%) Dominant species Coverage (%)
(Stem/ha)
1 Sophora japonica, pinus tabuliformis 81 404 Deutzia parviflora Bunge 61 Liriope graminifolia 40
2 Populus, sophora japonica 69 631 Weigela florida 53 Liriope graminifolia, Bambuseae 51
3 Populus 32 326 Lonicera maackii  44 Buchloe dactyloides 63
4 Salix babylonica 29 531 Ligustrum lucidum 23 Buchloe dactyloides 41
5 Populus 50 632 Sabina vulgaris Lonicera maackii 59 Liriope graminifolia 46
6 Sophora japonica 86 615 Syinga reticulata 23 Buchloe dactyloides 82
7 Platanus orientalis 23 319 Sabina vulgaris 27 Buchloe dactyloides 55
8 Platycladus orientalis 75 1,501 Lonicera maackii 43 Commelina communis 53
9 Ailanthus altissima, Sophora japonica 51 817 Lonicera maackii, Amygdalus triloba 39 Viola pekinensis 60
10 Pinus tabuliformis 22 401 Syinga reticulata 57 Poa annua, Viola pekinensis 40
11 Salix babylonica 59 322 Prunus cerasifera 54 Viola pekinensis 82
12 Acer truncatum bunge 47 336 Magnolia denudata 41 Poa annua 77
13 Acer truncatum bunge 92 636 Rosa xanthina 53 none 0
14 Pinus tabuliformis 12 229 Rosa chinensis 0 none 100
15 Fraxinus sogdiana Bunge 53 435 none 0 Buchloe dactyloides 100
16 Populus 49 863 none 0 Setaria viridis 100
17 Populus 65 1,006 none 0 Liriope graminifolia 100
18 Salix babylonica 44 869 none 0 Poa annua 24
19 Pinus tabuliformis 55 1,193 none 0 Poa annua 50
20 Populus, metasequoia glyptostroboides 51 1,229 none 0 Viola pekinensis 100
21 Pinus bungeana 4 645 none 61 Trifolium repens 21
22 Ulmus pumila, sabina chinensis 62 1,021 none 0 Viola pekinensis 40
23 Koelreuteria paniculate, pinus bungeana 78 996 none 0 Liriope graminifolia 39
24 Pinus tabuliformis, Pinus bungeana 82 1,100 none 0 Viola pekinensis 61
25 Populus 70 638 none 0 Viola pekinensis 55
26 Sophora japonica, salix babylonica 87 656 none 0 Viola pekinensis 31
27 Sophora japonica 93 673 none 0 Humulus japonicus 71
28 Sophora japonica 83 687 none 0 Poa annua 61
29 Populus 61 913 none 0 Liriope graminifolia 19
30 Salix babylonica 72 1,006 none 0 Potentilla chinensis, Viola pekinensis 37
31 Populus 70 961 none 0 Poa annua 95
32 Sophora japonica 68 811 none 0 Poa annua 79
33 Populus 82 863 none 0 Poa annua 66
34 Sabina chinensis 49 1,139 none 0 Poa annua 77
35 Koelreuteria paniculata, salix babylonica 78 996 none 0 Poa annua 91
36 Pinus tabuliformis 67 937 none 0 Poa annua 89
37 Ginkgo biloba 49 233 none 0 none 0
38 Sophora japonica 78 921 none 0 none 0
39 Koelreuteria 91 833 none 0 none 0
40 Sophora japonica, Fraxinus chinensis 83 563 none 0 none 0
41 Sophora japonica 95 1,114 none 0 none 0
42 Sabina chinensis 73 1,661 none 0 none 0
43 Cedrus deodara 55 812 none 0 none 0
44 none 0 0 Malus spectabilis,F. atropu tpurea 21 Poa annua 81
45 none 0 0 Prunus persica Chrysanthemoides 78 Poa annua 43
46 none 0 0 Amygdalus triloba 77 Poa annua 60
47 none 0 0 Magnolia denudata 23 Poa annua 95
48 none 0 0 none 0 Buchloe dactyloides 100
49 none 0 0 none 0 Poa annua 100
50 none 0 0 none 0 Poa annua, buchloe dactyloides 91

Note: “none” in the table means there is none or few of this type of plant (arbor or shrub or herb) in a sample plot.

 Vegetation quantity

There is a reasonable relationship between vegetation quantity and NDVI in pixel scale (Di et al., 2012) for the study area, and this was expressed as: V Q i = 1 ( 1 30000 + 0.0002 × 0.03 NDV I i ) Where VQ i (m2/900 m2) is the vegetation quantity in pixels, which were covered by sample plots, and NDVI i is the normalized difference vegetation index in pixels; the NDVI i estimation model was obtained from Eq. 2 (Di et al., 2012). NDV I i = ρ 5 − ρ 4 ρ 5 + ρ 4 Where ρ 5 is reflectivity in OLS-5 band, and ρ 4 is reflectivity in OLS-4 band.

Mean VQ per-plot was calculated as follows: V Q RS = 1 n ∑ n = 1 1 V Q i Where n is the number of pixels covered by a sample plot.

On 12 May 2013, 13 June 2013, and 1 Sep 2013, the VQ of 10 plots were measured with a LAI-2200 canopy analyzer (LI-COR, United States) to verify the accuracy of calculation of VQ RS .

 Land surface temperature

In this paper, The land surface temperature (T s) of different land-use types was calculated with a single-channel algorithm (Di et al., 2012), including road, urban water body, residential area, high vegetation quantity area and low vegetation quantity area, as shown in Eqs 4,5. The mean T s of a sample plot was expressed as an average of the temperature values of all the pixels covered by the sample plot in 3 days, which corresponded to days when the three remote sensing images were taken. These were defined by the equations below. T 10 = K 2 ln ( K 1 L 10 + 1 ) T S = T 10 ε 0.25 Where T 10 is radiant brightness temperature (K) in TIRS-10 band, °C, T s (K) is land surface temperature, °C, K 1 (m2 sr μm) and K 2 (K) are constants, ε is thermal emissivity, and L 10 is radiant brightness (m2 sr μm) in TIRS-10 band.

 Evapotranspiration

The calculation of evapotranspiration (ET) was based on SEBAL, which was established by (Bastiaanssen et al., 1998b) and widely used in the retrieval of ET by RS (Mohammad and Samaneh, 2012). The processes involved in SEBAL utilization can be summarized as the instantaneous latent heat fluxes of every pixel that can be gotten by surface energy balance, as expressed in equation Eq. 6. The instantaneous evaporative fraction is shown in the literature to be similar to the 24-h evaporative fraction (Shuttleworth et al., 1989), and thus daily average value of ET can be calculated through time scale transformation (Brutsaert and Sugita, 1992; Crago, 1996), as expressed in Eqs 7,8. The mean ET of a sample plot was expressed as an average of the ET values of all the pixels that were covered by the sample plot in 3 days, and this corresponded to days when the three RS images were taken. In Bajia Park, the actual ETs were measured in seven plots based on the principle of water balance to verify its accuracy. E T ins = R n − G − H λ A ins = R n − G − H R n − G = λET R n − G = A 24 E T 24 = R n 24 × Λ 24 × 86400 [ 2.501 − 0.002361 × ( T − 273.15 s ) ] × 10 6 where R n is the instantaneous net radiation (Wm−2); G is the instantaneous soil heat fluxes (Wm−2); H is the instantaneous sensible heat exchange between air and land surface (Wm−2); λ is the latent heat of vaporization of water (Wm−2 mm−1), and A ins and A 24 are the instantaneous and 24-h evaporative fractions. The algorithms of the above parameters are similar to those established by (Du et al., 2013).

 Model evaluation criteria

Using satellite data and meteorological data, the coefficient of determination (R 2), mean absolute relative error (MARE) and average relative error ( R ¯ e ) were used to measure the performance of the estimates of ET and VQ. In general, the R 2 value, which was calculated by linear regression analysis, is an indicator of the strength of relationship between the observed and simulated values. If the R 2 values are less than or very close to zero, the prediction of the modeling system is considered unacceptable or poor. The MARE and ( R ¯ e ) indicate the model’s ability to predict the values of a given prediction. They were defined as follows: M A R E = ∑ i = 1 n | x i ∗ / x i − 1 | n R ¯ e = ∑ i = 1 n ( x i ∗ − x i ) n × 100 % where x i is the ith observation, x i ∗ is the ith simulation value, and n denotes the total number of data points (observations) in the record. Lower values of MARE and ( R ¯ e ) are preferred.

 Evaluation of vegetation eco-water-saving ability

A comprehensive eco-water-saving ability index of vegetation (EWI) was proposed to evaluate vegetation cooling and water-saving abilities of different vegetation structures. EWI can be defined as: EW I i = 1 − T i T ¯ × E T i / LA I i ET / LAI Where EWI i is the comprehensive eco-water-saving ability index value of the ith vegetation structure, and it is dimensionless value greater than 0; T i is the land temperature of the ith vegetation structure, °C; T ¯ is the mean land temperature of all 50 sample plots in °C; LAI is leaf area index, and both LAI and VQ can reflect the leaf volume of plants. ET i/LAI i is the ith daily evapotranspiration per unit LAI of the ith vegetation structure in mm/d; and ET / LAI is the mean daily evapotranspiration per unit LAI of all 50 sample plots in mm/d. A high EWI indicates that the comprehensive eco-water-saving ability of the vegetation is preferred.

 Result Surface temperature of the study area

The results showed that temperature between different land use types in the growing season varies (Figure 2). The T s of the road was the highest, followed by those of the residential area, low vegetation quantity area, high vegetation quantity area, and urban water body, respectively. Compared with other land use types, the temperatures of the urban water body and high vegetation quantity area are 9–18 degrees lower. This indicates that vegetation can help to relieve the urban heat island effect, and vegetation index appears to be an indicator of environmental temperature (Gallo et al., 1993).

Dynamics of Ts patterns in the study area in (A) May and (B) June 2013 based on the surface energy balance method using Landsat-8 satellite data. Each date had different ranges.

 Evapotranspiration of study area

After an evaluation of the ET results calculated using the SEBAL model and that measured based on the water balance principle of four sample plots on May-12-2013, June-13-2013, and Sep-1-2013, a linear relationship was observed between the calculated ET and measured ET; R 2 was 0.445, MARE was 0.16, and ( R ¯ e ) was −21.3%. This indicated that the accuracy of the model calculation is acceptable. Errors may arise due to interference in the pixel data of sample plots by land objects in the surroundings of these sample plots, such as buildings and roads.

The distribution of ET on May-12-2013 and June-13-2013 is shown in Figure 3; the highest ET was observed in water bodies such as the Kunming Lake, Fuhai, Jing-Mi water diversion canal, and Olympic lake (7.2 ± 1.6 mm). The second highest ET value was observed in places with greater vegetation coverage. For instance, the Olympic Forest Park, the summer Palace, and Dongsheng country parks were in the area bounded by the dotted line (4.3 ± 2.1 mm). The third-highest ET regions were the residential areas and greenbelts along the city roads with lower vegetation coverage (3.3 ± 1.7 mm). Dense human settlements and commercial districts had the lowest ET value which shows that vegetation played a significant role in the water consumption of the soil-plant-atmosphere system. In different months, ET in June is lower than that in May, because May is the dry season, with less rainfall and less air humidity, while June is the rainy season, with more rainfall and higher air humidity, resulting in smaller ET.

Dynamics of evapotranspiration patterns in the study area during (A) May and (B) June, 2013 based on surface energy balance method using Landsat-8 satellite data. Each date had different ranges.

 Vegetation quantity of study area

Based on an evaluation of VQ results calculated using the NDVI and actual VQ observation of 12 sample plots on May-12-2013, June-13-2013, and Sep-1-2013, a linear relationship was detected between the calculated VQ and observed VQ; R 2 was 0.66, MARE was 0.21, and ( R ¯ e ) was 13.5%, which indicated that the accuracy This of the model calculation is acceptable. The main sources of error resource may be similar to that of the ET inversion. The mean vegetation quality of the study area is presented in Figure 4; the green land area is 56.87 km2, which is 38.6% of the total study area. Vegetation quality area is 822.1 km2.

Distribution of vegetation quality per pixel; the area of a single pixel is 900 m2.

 Relationships between surface temperature, evapotranspiration, and <italic>LAI</italic>

The mean T s, ET, and LAI of 50 sample plots were presented in a scatter plot (Figure 5). A negative correlation was observed between mean T s and mean LAI; an increase in mean LAI by 1, can reduce T s by 0.62-degree centigrade. However, a positive correlation was observed between mean ET and mean LAI; an increase in mean LAI increase by 1 mm/d can increase mean ET by 0.073 mm/d.

Relationships between mean T s, mean ET and mean LAI. Regression line for mean ET and mean LAI can be expressed as: ET = 3e 0.019LAI, R 2 = 0.127; p < 0.01 (unfilled circles, unbroken line). Regression line for mean Ts and mean LAI can be expressed as: ET = −0.62LAI + 41.42, R 2 = 0.49; p < 0.01 (filled circles, broken line).

To analyze the effect of species, the mean ET of sample plots were expressed as ET per LAI (ET/LAI), and the relationship between mean T s and ET/LAI was determined as Figure 6. There were 11 species (indicated by the points in a solid line ellipse) with a mean T s < 30°C. Their corresponding serial numbers are 8, 9, 10, 16, 22, 23, 24, 25, 29, 45, and 48 (Table 1), respectively, accounting for 22% of the total. Eleven species (indicated by points in a dashed ellipse) had ET/LAI < 0.23 mm/d, accounting for 22% of the total. Their corresponding serial numbers are 1, 2, 11, 12, 13, 16, 32, 39, 40, 44, and 47, respectively (Table 1).

Relationships between T s and ET/LAI.

As shown in Table 2, the EWIs of species or species compositions in 50 sample plots were divided into three categories based on the natural breakpoint method by their EWI value: Good (EWI ≥ 0.795), Common (0.795 > EWI ≥ 0.419) and Weak (EWI < 0.419). Each category included 5–29 species or species compositions. The species or species compositions were classified as Good, it means that the species or species need less water for per unit LAI.

EWI of species or species compositions of 50 example plots.
Good Serial number 39 24 9 47 2 16 8 40 29 12 11 48 38 22 10
EWI 0.99 0.985 0.984 0.975 0.957 0.956 0.956 0.955 0.955 0.951 0.946 0.943 0.942 0.936 0.934
Good Serial number 44 45 17 13 23 32 25 19 1 46 5 49 18 26
EWI 0.924 0.918 0.917 0.912 0.899 0.899 0.885 0.88 0.86 0.856 0.825 0.824 0.803 0.795
Common Serial number 27 4 20 30 50 14 21 43 33 15 31 7 36 41
EWI 0.762 0.734 0.734 0.688 0.683 0.674 0.65 0.64 0.61 0.603 0.553 0.496 0.443 0.419
Weak Serial number 6 3 42 37 28 34 35
EWI 0.376 0.21 0.192 0.172 0.118 0.098 0

Note: Serial numbers here corresponds to the serial numbers in Table 1 of 50 sample plots.

 Discussion

Two characters of vegetation structures with lower Ts can be found: 1) those with higher arbor density, 2) warm-season turf grasses. This because the arbor has higher vegetation quantity and more water is used in cooling air by transpiration compared to that used by shrubs and herbs. Furthermore, the evapotranspiration of warm-season grasses is higher than that of cool-season grasses (Aronson et al., 1987; Xiao et al., 2006). The vegetation structures with lower ET can be summarized as 1) thin native trees with greater species composition, such as Sophora japonica and Pinus tabuliformis, 2) cool-season turf grasses. Since the average water consumption of native vegetation is lower than that of extrinsic vegetation in arid areas, it is difficult to have optimal cooling and water-saving abilities for green spaces. However, because of water shortages, green space managers in Semi-humid areas have to compromise with ecological benefit for less irrigation. In that case, species or species compositions which can balance ecological benefit and water-saving for urban green space should be used.

In this study, 50 typical vegetation structures were divided into three categories according to EWI; the main difference can be found by a comparison of the different categories: the proportion of vegetation structures which include the arbor, shrub, and herb is 28% in the category with good comprehensive eco-water-saving ability; and the proportion is 21% and 29%, respectively, for the remaining two categories. In addition, the proportion of vegetation structures without shrub is 79% and 71% in the categories with common and weak eco-water-saving ability, respectively, and these are significantly higher than that of the category Good. It is indicated that, generally, the arbor-shrub-herb mix and complete coverage by shrubs gave vegetation structures a better comprehensive eco-water-saving ability. This ability can be explained by the features of their vegetation structures such as great heat capacity and high reflectivity, and relatively low water consumption (Zhou et al., 2005; Song et al., 2007).

 Conclusion

Urban ecological environment and water resources are key issues of concern. In this study, RS inversion methods were utilized to calculate the land temperature, evapotranspiration and vegetation quantity of 50 typical green space vegetation structures. According to verification tests, the accuracy of calculation is acceptable. The relationships between Ts, VQ, and ET of green land were determined with regression analysis. An eco-water-saving abilities index (EWI) was proposed to evaluate different green space structures. Fifty species or species compositions were classified into good (EWI ≥ 0.795), common (0.795 > EWI ≥ 0.419) and weak (EWI < 0.419) categories, respectively, based on the assessment results. The characteristics of 17 species or species compositions with good cooling and water-saving abilities were summarized as 1) a mix of arbor, shrub and herb, 2) complete coverage of green space by shrub. The method and eco-water-saving species or species composition can be utilized in developing possible models of urban green space management, and were of significance in improving the ecological carrying capacity of the Semi-humid Region.

 Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

 Author contributions

XL: Conceptualization, Methodology, Formal analysis, Writing—Original Draft, Writing—Review and Editing, Visualization. YL: Methodology, Formal analysis, Writing—Original Draft, Software, Visualization. SD: Resources, Data Curation, Writing—Review and Editing, Supervision, Project administration, Funding acquisition. YN: Conceptualization, Methodology, Writing—Review and Editing. CZ: Conceptualization, Methodology, Visualization.

 Funding

This study was financially supported by the Natural Science Foundation of Shandong Province (Grant No. ZR2016DB12).

The insightful comments by editors and reviewers are greatly acknowledged. These comments helped us improve our original manuscript greatly.

 Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

 Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

 References Aronson L. J. Gold A. J. Hull R. J. Cisar J. L. (1987). Evapotranspiration of cool‐season turfgrasses in the humid northeast 1 . Agron. J. 79, 901–905. 10.2134/agronj1987.00021962007900050029x Bastiaanssen W. G. M. Menenti M. Feddes R. A. Holtslag A. A. M. (1998). A remote sensing surface energy balance algorithm for land (SEBAL). 1. Formulation. J. Hydrol. X. 212, 198–212. 10.1016/s0022-1694(98)00253-4 Bastiaanssen W. G. M. Pelgrum H. Wang J. Ma Y. Wal T. V. D. Roerink G. (1998). A remote sensing surface energy balance algorithm for land (SEBAL). J. Hydrol. X. 212-213, 213–229. 10.1016/s0022-1694(98)00254-6 Becker F. (1987). The impact of spectral emissivity on the measurement of land surface temperature from a satellite. Int. J. Remote Sens. 8, 1509–1522. 10.1080/01431168708954793 Beijing Municipal Bureau of Statistics (2019). Beijing statistical yearbook. Available at: http://www.bjstats.gov.cn/tjnj (accessed June, 26) Brutsaert W. Sugita M. (1992). Application of self-preservation in the diurnal evolution of the surface energy budget to determine daily evaporation. J. Geophys. Res. 97, 18377–18382. 10.1029/92jd00255 Cai X. (2008). Water stress, water transfer and social equity in northern China—implications for policy reforms. J. Environ. Manage. 87, 14–25. 10.1016/j.jenvman.2006.12.046 Cheng X. Huang M. Shao M. Warrington D. N. (2009). A comparison of fine root distribution and water consumption of mature Caragana korshinkii Kom grown in two soils in a semiarid region, China. Plant Soil 315, 149–161. 10.1007/s11104-008-9739-5 Crago R. D. (1996). Conservation and variability of the evaporative fraction during the daytime. J. Hydrol. X. 180, 173–194. 10.1016/0022-1694(95)02903-6 Dash P. Göttsche F. M. Olesen F. S. Fischer H. (2002). Land surface temperature and emissivity estimation from passive sensor data: Theory and practice-current trends. Int. J. Remote Sens. 23, 2563–2594. 10.1080/01431160110115041 Di S. C. Wu W. Y. Liu H. L. Pan X. Y. (2012). Water consumption estimation and evapotranspiration inversion based on remote sensing technology. Trans. Chin. Soc. Agric. Eng. 28, 98–104. Du J. Song K. Wang Z. Zhang B. Liu D. (2013). Evapotranspiration estimation based on MODIS products and surface energy balance algorithms for land (SEBAL) model in Sanjiang Plain, Northeast China. Chin. Geogr. Sci. 23, 73–91. 10.1007/s11769-013-0587-8 Gallo K. P. Mcnab A. L. Karl T. R. Brown J. F. Hood J. J. Tarpley J. D. (1993). The use of a vegetation index for assessment of the urban heat island effect. Int. J. Remote Sens. 14, 2223–2230. 10.1080/01431169308954031 Granger R. J. Gray D. M. (1989). Evaporation from natural nonsaturated surfaces. J. Hydrol. X. 111, 21–29. 10.1016/0022-1694(89)90249-7 Hobbins M. T. Ramírez J. A. Brown T. C. (1999). The complementary relationship in regional evapotranspiration: The CRAE model and the advection-aridity approach. Proc. Ninet. Annu. A. G. U. Hydrol. Days, 199–212. Idso S. B. Jackson R. D. Reginato R. J. (1975). Estimating evaporation: A technique adaptable to remote sensing. Science 189, 991–992. 10.1126/science.189.4207.991 Jackson R. D. Reginato R. J. Idso S. B. (1977). Wheat canopy temperature: A practical tool for evaluating water requirements. Water Resour. Res. 13, 651–656. 10.1029/wr013i003p00651 Jiménez Muñoz J. C. Sobrino J. A. (2003). A generalized single‐channel method for retrieving land surface temperature from remote sensing data. J. Geophys. Res. 108 (D22), 2003JD003480. 10.1029/2003jd003480 Katherine N. I. Richard A. F. Patrick D. W. Jamie T. Sarah R. P. Philip H. W. (2010). Ecological and psychological value of urban green space. Dimensions Sustain. City Future City 2, 215–237. Katri O. Aigars I. Iveta V. K. Malle M. Tatjana K. Jaan K. (2011). Changes in the canopies of Pinus sylvestris and Picea abies under alkaline dust impact in the industrial region of Northeast Estonia. For. Ecol. Manage. 262, 82–87. 10.1016/j.foreco.2010.07.031 Kong F. Nakagoshi N. Yin H. Kikuchi A. (2005). Spatial gradient analysis of urban green spaces combined with landscape metrics in Jinan City of China. Chin. Geogr. Sci. 15, 254–261. 10.1007/s11769-005-0038-2 Li Z. Z. Li G. H. Zhao B. X. Lai N. N. Jing L. J. (2009). Water consumption of green space and gardens in the Summer Palace of Beijing. J. Beijing For. Univ. 31, 66–72. Liu C. S. Gao Z. Q. Gao W. (2007). Retrieval evapotranspiration and land surface temperature in response to land use/cover change based on remote sensing data. Trans. Chin. Soc. Agric. Eng. 23, 1–8. Liu L. Guan D. Peart M. R. Wang G. Zhang H. Li Z. (2013). The dust retention capacities of urban vegetation—A case study of guangzhou, south China. Environ. Sci. Pollut. Res. 20, 6601–6610. 10.1007/s11356-013-1648-3 Masoud E. Sayyed A. K. Ali A. M. Ali B. (2007). Water consumption of a six-year-old river red gum plantation in the Southern Zagros Mountains, Iran. J. Mt. Sci. 4, 136–145. 10.1007/s11629-007-0136-y Mohammad T. D. Samaneh P. (2012). Evaluation of water balance in a mountainous upland catchment using SEBAL approach. Water Resour. manage. 26, 2069–2080. 10.1007/s11269-012-9999-y Perino G. Andrews B. Kontoleon A. Bateman I. (2014). The value of urban green space in britain: A methodological framework for spatially referenced benefit transfer. Environ. Resour. Econ. (Dordr). 57, 251–272. 10.1007/s10640-013-9665-8 Qin Z. Karnieli A. Berliner P. (2010). A mono-window algorithm for retrieving land surface temperature from Landsat TM data and its application to the Israel-Egypt border region. Int. J. Remote Sens. 22, 3719–3746. 10.1080/01431160010006971 Roerink G. J. Su Z. Menenti M. (2000). S-SEBI: A simple remote sensing algorithm to estimate the surface energy balance. Phys. Chem. Earth Part B Hydrology Oceans Atmos. 25, 147–157. 10.1016/s1464-1909(99)00128-8 Shuttleworth W. J. Gurney R. J. Hsu A. Y. Ormsby J. P. (1989). The variation in energy partitioning at surface flux sites, remote sensing and large scale global processes. Wallingford, UK: IAHS. Song Y. T. Yu S. X. Li N. (2007). Spatial structure of the surface temperature in Shenzhen. Acta Ecol. Sin. 27, 1489–1498. Su Z. (2002). The Surface Energy Balance System (SEBS) for estimation of turbulent heat fluxes. Hydrol. Earth Syst. Sci. 6, 85–100. 10.5194/hess-6-85-2002 The National Afforestation Committee (2019). China greening bulletin. Available at: http://www.forestry.gov.cn//portal/main/s/195/content -531395.html (accessed July, 12). Tohru M. Keitaro I. Daisuke H. Dai I. Takashi U. Shuji I. (2011). Landscape ecology in asian cultures. Japan: Springer. World Resources Institute (2005). Millennium ecosystem Assessment: Ecosystems and human well-being: Biodiversity synthesis. Washington DC: IslandPress. Xiao J. Sun X. H. Li G. Y. (2006). The water consumption regularity and irrigation system of Turfgrass under the condition of spray irrigation. Sci-Tech Inf. Dev. Econ., 152–153. Yan M. Yamanaka N. Yamamoto F. Du S. (2010). Responses of leaf gas exchange, water relations, and water consumption in seedlings of four semiarid tree species to soil drying. Acta Physiol. Plant. 32, 183–189. 10.1007/s11738-009-0397-x Yen T. Lee J. (2011). Comparing aboveground carbon sequestration between moso bamboo (Phyllostachys heterocycla) and China fir (Cunninghamia lanceolata) forests based on the allometric model. For. Ecol. Manage. 261, 995–1002. 10.1016/j.foreco.2010.12.015 Zhao K. L. Fu W. J. Jiang P. K. Zhou G. M. (2013). Functions of natural organic matter in changing environment. Zhejiang (China): Zhejiang University Press. Zhao W. Niu Z. Chang X. Li S. (2010). Water consumption in artificial desert oasis based on net primary productivity. Sci. China Earth Sci. 53, 1358–1364. 10.1007/s11430-010-4028-6 Zhou L. C. Shi W. Y. Xue W. J. Wang T. Z. Ge Z. M. Zhou H. (2005). Relationship between vegetation structure and the temperature and moisture in urban green spaces of Shanghai. Chin. J. Ecol. 24, 1102–1105.