基于卫星数据分析云南火电厂对地区水质的影响（英文版）.pdf

Understanding the Effects of Thermal Power Plants on Regional Water Quality Based on Satellite-derived Data in Yunnan Planet Data Inc. 2020.12.301 I Background With the rapid growth of economy, the demand of electricity consumption for daily production and peoples daily life is exponentially increasing. Thermal power generation is still the main mode in China. In 2017, The total power generation reaches 6.5 trillion kwh. Among them, thermal power plants account for more than 71% of power generation. In addition to air, water is usually chosen as the cooling medium. In 2015, the Standing Committee of the Political Bureau of the Central Committee deliberates and approves the Action Plan for the Prevention and Control of Water Pollution (Water 10), aiming to strengthen efforts to prevent and control water pollution and ensure national water security. The plan calls for more than 70 percent of seven river basins, including the Yangtze River and Zhuhai, to have water that reaches or exceeds quality category III by 2020. Yunnan province is located at the confluence of the Yangtze River basin and the Pearl River Basin. Although the basin area of nine plateau lakes represented by Yangzonghai and Dianchi lakes only accounts for 2.1% in the area of Yunnan Province, they play an important role in the economic and social development of Yunnan province and account for more than one-third of the provincial GDP annually. In recent years, as the industry and tourism develop, Yangzonghai and Dianchi witness a rapid degradation in water quality. Yunnan provincial government and Provincial Environmental Protection Bureau attach great attention to the comprehensive control of water pollution in Yangzonghai and Dianchi and treat it as a major implement of sustainable development. However, due to the influence of environmental factors, water exchange periods of two lakes are long and the ecosystems are fragile. The high-density population brought by the developed economy aggravates the load of pollution in lakes. The multifarious types of pollutants interact with each other, which increases the difficulty of implementing water quality management and leads to long-term accumulation of pollutants. Due to the intense urbanization, the buffer zone is underdeveloped, which leads to fragile plant communities and the decline of self-purification capacity of lakes. Therefore, discharge control is critical in improving the water quality of Yangzonghai and Dianchi and passing the window period of ecological restoration.2 To identify illegal discharges from enterprises and individuals, monitor the discharge of major pollution source and provide theoretical and data support for further development of pollution control policies, a detailed evaluation of the temporal and spatial changes in water temperature, chlorophyll concentration and water transparency is required. However, due to the cost and difficulty of implementing field trips, the data obtained have limited coverage in time and space, which will lead to a failure in carrying out long-term spatial and temporal analysis of Yangzonghai and Dianchi. With the development of remote sensing technology, methods of using empirical formula to evaluate water quality has been optimized and popularized by researchers. Compared to the traditional sampling methods, environmental remote sensing technology has higher spatial and temporal resolution. To investigate the relationship among water temperature, transparency, and chlorophyll concentration, and their temporal ad spatial variations, we develop three models to respectively represent historical situations. Taking advantage of the satellite-based models, we analyze the long-term trend and whole scale variations. Hypotheses are proposed that discharged warm water could deteriorate aquatic ecosystem by providing a warm and nutrient-rich environment for plankton and plants. Previous studies have pointed out that the discharged warm water from thermal power plants may cause multiple kinds of damage to water quality, including increasing the water temperature (thermal pollution), increasing the concentration of suspended particulate matter, and changing chemical composition of water body. However, the remote sensing technology mainly works for revealing correlations, rather than casual relationships, the intermediate processes are not discussed in this report. Analyses are merely based on satellite-derived observations. We ignore the principles of the intermediate reactions among pollutants and aquatic organisms. For instance, the way that discharged warm water impacts the vigor of plankton and plant3 II Method 1. Study Area 1) Yangzonghai Yangzonghai, located in the southeast of Kunming, covers an area of 31.9 square kilometers, with an average water depth of 20 meters. It stores 604 million cubic meters of water, which is about half of the water volume of Dianchi Lake. According to the 2015 Environmental status Bulletin of Yunnan Province, the water quality of Yangzonghai in 2015 is classified as Class IV. Arsenic concentration is classified as Class IV, which is 0.05 times above the standard. Phosphorus and chemical oxygen are 0.36 times and 0.17 times above the standard respectively. The average nutritional status index of the whole lake is 41.2, which is classified as mesoeutrophic. 2) Dianchi Dianchi is the largest lake in southwest China, belonging to the Yangtze River Basin. It is in the south-central part of Kunming Basin. The lake covers an area of 300 square kilometers and the shoreline is about 150 kilometers long. In the north of the lake, there is an embankment stretching from east to west, which is 3.5 kilometers long and 300 meters wide. It divides Dianchi into two parts. South of the embankment, known as the outer sea, is the main part of the Dianchi, covering an area of 289.065 square kilometers, accounting for 97.2% of the total area. North of the embankment is called inner sea, which is also known as grass sea, occupying an area of about 10 square kilometers. The average depth of Dianchi is about 5 meters. 2. Data acquirement and pre-processing In order to obtain sufficient data for time series analysis, we expand the time period to 2006 to 2018. For the data pre-processing, we firstly conduct radiometric calibration for Landsat5TM and Landsat8OLI images, and the calibration type is radiometric brightness. Secondly, we calibrate the images based on sensor types, the parameters acquired by each image (season, aerosol model, atmospheric model, visibility, etc.), the altitude and regional type of the study area. Thirdly we use the object-oriented image segmentation technology to extract the vector boundary of Yangzonghai and Dianchi Lake. Then, we utilize the vector boundary as a mask4 to extract areas of Yangzonghai and Dianchi Lake. Finally, for areas covered by clouds, we remove the noise based on spectrum signature. 3. Modeling 1) Water surface temperature model The inversion of water surface temperature is based on atmospheric correction method. The expression for the infrared thermal luminance value ! ! received by the satellite sensor: != !(!)+(1 !)! !+! （ 1） Where, is the surface emissivity, ! is the true temperature (K), !(!) 为 is the luminance of the blackbody at temperature T, is the atmosp heric transmittance in thermal infrared band, ! is upward atmospheric luminance, and ! is downward atmospheric luminance. The luminance of the blackbody at temperature T !(!) is expressed as following: !(!)=! ! !(1 !)!/ （ 2） ! is calculated by Planck formula: !=! 2 /ln(! 1 /!(!)+1) （ 3） 2) Water transparency model The change of water transparency (SD) is mainly affected by the optical components (algae, non-algal particles, yellow substances). Transparency is also an important index to evaluate eutrophication, which directly reflects the clarity and turbidity degree of the lake. The reflectance of red and near-infrared bands is easily affected by suspended matters. Suspended matters have a strong negative correlation with transparency. However, it is rarely used due to the strong absorption of near-infrared bands in water. Based on the reflectance characteristics of each band, we select the ratio of red and green bands to construct a water transparency model: Ln(SD)= a * ( B Green /BRed ) b (4) 3) Chlorophyll concentration model Based on the empirical models for the same water area, the ratio of near-infrared band to visible red band is used as a sub-factor, which can effectively minimize the influence of the atmospheric effect. The model established is the relationship between the natural logarithm of5 chlorophyll A (chla) and B NIR /B red : Ln(chla)= a - b/(B NIR /Bred ) （ 5） 4) Evaluating model According to the requirements from Chinas environmental monitoring station surface water environmental quality assessment method, the evaluation of surface water quality is referenced by standard GB3838-2002. Lakes and reservoirs nutritional status evaluation indexes include the chlorophyll a (chla), total phosphorus (TP), total nitrogen (TN), transparency (SD) and potassium permanganate index (CODMn). The nutrition level rangers from 1 to 5 which is from poor to severe eutrophication. TLI（ chla ） = 25 + 10.86lnchla (6) TLI（ TP ） = 94.36 + 16.24lnTP (7) TLI（ TN ） = 54.53 + 16.94lnTN (8) TLI（ SD ） = 51.18 - 19.4lnSD (9) TLI（ CODMn ） = 1.09 + 26.61lnCO DMn (10) Where the unit of chla is mg/m 3 , and the unit of SD is m. Units of other indicators are mg/L. In this study, instead of chlorophyll concentration and water transparency, TLI(chla) and TLI(SD) are used as water quality evaluation indexes.6 III Result 1. Yangzonghai 1 ）Water temperature Figure 1. Yangzonghai land uses (left) and research area division (right) The land uses along the Yangzonghai coast are diverse (Figure 1), mainly including towns, woodland, and paddy fields. The south bank is majorly occupied by agricultural land. The land uses in the west bank include towns, cities, woodland orchards, and mining. According to the Manual of Discharge Coefficient of Livestock and Poultry Industry, the pollution along the southwest shoreline of Yangzonghai mainly comes from livestock breeding. The northern part is a cluster of towns and industries. Based on the statistics from the second national pollution census, industrial pollution in the Yangzonghai Basin mainly comes from the northern part, accounting for 63.1% of the total discharges in 2018. Nine companies, including the Yangzonghai Power Plant, discharge 362,100 tons of wastewater. According to the above information, we divide the waters of Yangzonghai into southern, central, and northern regions to distinguish the effects of various land uses.7 Figure 2. Yangzonghai maximum (left) and average (right) water temperature in winter Due to the limitations of Landsat temporal resolution and climatic conditions, as well as the consideration of amplifying the impact of agricultural water, residential water and discharged warm water, data from winter (December, January, and February) data are mainly selected for analysis. In terms of spatial distribution, the highest water temperature is found in the river channels and coastal areas dominated by urban. We assume that the highest temperature represents the temperature of pollutants (including water itself), and the location where they appear is the source of pollution. Variations in average temperatures in different regions indicate that there is a significant feature in the spatial distribution of the temperature in Yangzonghai. From 2006 to 2018, the highest water temperature in the northern Yangzonghai is always higher than that in the central and southern part, as well as the average water temperature. The average temperature in the north, central and south of the 12 years is 13.6, 13.1 and 13.0, respectively. Therefore, out preliminary analysis suggests that the main source of the overall temperature rise of Yangzonghai is industrial cluster represented by the Yangzonghai Power Plant in the northern part. In order to investigate the influence on the water temperature from outside, we visualize the water temperature by taking the temperature of unaffected water as the reference temperature. There are two main methods to calculate the reference temperature. The first is the average temperature obtained from the core area that is not affected by the outside. The second is to take the average water temperature of the whole research area as a reference, and then calculate the average temperature after excluding the areas that have higher temperature than the average. Regions above the base temperature is known as the warming zones. We determine the reference temperature by referring the second method.8 Figure 3. Distribution of Yangzonghai warming zones in 2009 (left) and distribution of Yangzonghai North Coast warming zones (right) After calculating the reference temperature and visualizing warming zones, the results (Figure 3) are consistent with the hypothesis. The southwest coast is affected by the pollution from large-scale livestock breeding. The water temperature is much higher than the reference temperature. However, due to the average depth of Yangzonghai, the discharged water cannot threaten the core area, and the warming area is zonally distributed along the southwest coast. Different from the southwest coast, the discharged warm water from the north coast enters the Yangzonghai through channels near the Yangzonghai power plant, and the warming zone extends forward the core area in a radiative pattern. Figure 4. Distribution of warming zones Since the water temperature in winter is the lowest all year round, and the temperature9 difference between discharged warm water and the natural water body is the largest, the distribution of warming zones can be more clearly observed. Although other land uses in the northern part, such as parks, agriculture, and tourism, also impact water temperature, data and images over the years show that discharged warm water from channels near the Yangzonghai Power Plant contributes significantly on water temperature in the northern Yangzonghai. Figure 5. Area proportion of Yangzonghai warming zones in spring (left) and winter (right) Figure 6. Area of Yangzonghai warming zones in spring and winter Figure 5 shows the proportion of Yangzonghai warming zones in spring and winter, as well as the trend. Each bar represents the percentage of the area above the reference temperature. The orange and gray parts in the column respectively represents the proportion of the areas that are 1 and 2 higher than the reference temperature. From figure 6, we find that during 2006-2007, the area of spring warming zone is significantly larger than that of winter, with a difference of nearly 10km 2 .10 2 ）TLI （chla ）and TLI （SD ） Figure 7. Spatial distribution of water surface temperature (left), TLI(chla) (middle) and TLI(SD) (right) in northern Yangzonghai in 2017 Figure 8.