Method for simulating spatial and temporal change of drainage basin runoff under climate change based on RCCC-WBM model

Abstract

The invention discloses a method for simulating spatial and temporal change of drainage basin runoff under climate change based on an RCCC-WBM model. The method comprises the following steps: acquiring actually measured rainfall, runoff, air temperature, evapotranspiration data and future climate data of a research area; performing cutting, interpolation, downscaling and multi-mode averaging processing on the corresponding future climate data, and calculating corresponding future evapotranspiration data; constructing a model, and performing parameter calibration and verification; 4 climate estimation scenes are classified and input into the model, and future runoff is simulated; and performing time change analysis on the monthly future runoff sequence. According to the method, the problems that a hydrological model adopted by an existing method is complex in structure, multiple in parameter, difficult to calibrate, insufficient in resolution ratio, low in precision and different in regional adaptability are fully solved; according to the method, a linear trend analysis method and a nonlinear trend analysis method are combined, the future runoff trend is analyzed, and the problem that an existing prediction method is not comprehensive enough in time change characteristic analysis is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrological forecasting, and in particular to a method for simulating temporal and spatial variation of basin runoff under climate change based on an RCCC-WBM model. Background Art

[0002] Climate change is considered to be the main factor affecting surface water resources. Global climate models are effective tools for studying climate change mechanisms and estimating future climate conditions. GCMs have different initial conditions, parameterizations, spatiotemporal resolutions, and boundary conditions, which affect the output of the model. River basin hydrological models are important tools for understanding regional hydrological cycle processes, conducting hydrological forecasts, and managing water resources. Runoff and water volume simulations are indispensable and important aspects of flood control and disaster reduction, reservoir scheduling, water resources development and utilization, and environmental change impact assessment.

[0003] Conducting basin runoff change test analysis research has important practical value and scientific significance for effectively supporting basin development. However, the current methods for predicting future runoff changes still have the following defects: First, the hydrological models for most existing basin runoff simulations are complex in structure, have many parameters, and are difficult to calibrate, which greatly affects the accuracy of the prediction results; second, most existing hydrological models have insufficient resolution and cannot provide high-resolution spatial and temporal predictions; third, most existing hydrological models lack local details and do not provide detailed descriptions of small basins or specific areas; fourth, most existing hydrological models have low accuracy and have different regional adaptability; fifth, when analyzing the future trend of runoff, linear trend changes are mostly used, and the analysis of temporal change characteristics is not comprehensive enough. Summary of the invention

[0004] In order to realize the prediction of temporal and spatial variation of watershed runoff under climate change, the present invention provides a method for simulating the temporal and spatial variation of watershed runoff under climate change based on RCCC-WBM model.

[0005] A test method for the spatiotemporal variation of basin runoff under climate change based on the water balance formula is proposed. First, the RCCC-WBM model is constructed based on the measured monthly precipitation data, monthly temperature data, and monthly evapotranspiration data. Secondly, the future climate data simulated by the future climate model under different scenarios are combined to calculate the future evapotranspiration data. Then, the future precipitation, temperature, and evapotranspiration data under different scenarios are input into the model to obtain the future runoff sequence of the basin under climate change. Finally, the trend analysis of future runoff is carried out.

[0006] The present invention specifically adopts the following technical solutions:

[0007] A method for simulating the spatiotemporal variation of watershed runoff under climate change based on the RCCC-WBM model includes the following steps:

[0008] S1. Obtain the measured precipitation, runoff, temperature, and evapotranspiration data of the study area, including monthly average precipitation, monthly runoff, monthly average temperature, monthly average evapotranspiration, watershed area, and future climate data; future climate data include future monthly precipitation, maximum temperature, minimum temperature, and average temperature, and pre-process the data;

[0009] S2. The corresponding future climate data in step S1 are clipped, interpolated, downscaled, and multi-model averaged, and the corresponding future evapotranspiration data are calculated using the Hargreaves formula;

[0010] S3. Based on the measured precipitation, runoff, temperature and evapotranspiration data corresponding to step S1, the RCCC-WBM model based on water balance is constructed, and the measured monthly runoff data of the corresponding stations in the study area are used to calibrate and verify the model parameters;

[0011] S4. Classify the future climate monthly data processed in step S2 into four climate prediction scenarios, input them into the RCCC-WBM model established in step S3, and simulate the future runoff;

[0012] S5. Use seasonal index, Mann-Kendall linear trend and nonlinear trend test methods to analyze the temporal changes of the monthly future runoff series obtained in step S4.

[0013] Furthermore, the preprocessing of meteorological data in step S1 includes:

[0014] Step 11, performing missing value test, consistency test and white noise processing on the measured precipitation, temperature and evapotranspiration data. The specific method belongs to the existing technology and will not be described here.

[0015] Step 12: interpolate and extend the measured precipitation, temperature, and evapotranspiration data, and convert the units of the measured monthly runoff data to 3 ) is converted into runoff depth units (mm).

[0016] Furthermore, the step S2 comprises:

[0017] Step S21, cutting the acquired future climate data to the corresponding research area;

[0018] Step S22, using bilinear interpolation to interpolate all variables in the 19 climate models in CMIP6 into a grid so that the spatial resolution is consistent with the meteorological driving dataset;

[0019] Step S23, using the Delta downscaling method to downscale the climate data of 19 climate models in CMIP6 under four projection scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5). In describing the change of climate elements, precipitation uses the change rate, while temperature uses the change amount. The calculation formula for future precipitation and temperature scenarios of meteorological stations is:

[0020]

[0021] T f =T o +(T Gf -T Go ) (2)

[0022] Where: P f , T f is the future precipitation and temperature series reconstructed by the Delta method; P Gf , T Gf is the future precipitation and temperature series predicted by the climate model; P Go , T Go It is the multi-year average precipitation and temperature simulated by the climate model during the reference period; P0 and T0 are the multi-year average precipitation and temperature of the observation field during the reference period.

[0023] Step 24, calculating the multi-model average of future climate data. The future climate data after downscaling analysis is averaged according to multiple models to obtain multi-model average future climate data.

[0024] Here, the average values ​​of 19 climate modes for monthly precipitation, the average values ​​of 19 climate modes for maximum temperature, the average values ​​of 19 climate modes for minimum temperature, and the average values ​​of 19 climate modes for average temperature are calculated respectively.

[0025] Step S25, calculating future evapotranspiration data. The future climate data averaged by multiple models is calculated using the Hargreaves formula to obtain future evapotranspiration data. Hargreaves formula:

[0026] P=a×R a (T mean +17.8)(T max -T min ) b (3)

[0027] Where: a and b are empirical coefficients, with common values ​​of 0.0023 and 0.5; P is the potential evapotranspiration calculated by this method (mm / month); R a is the annual average monthly radiation (mm / month); T mean is the monthly average temperature, ℃; T maxis the highest temperature of the month, ℃; T min The lowest temperature of the month, ℃.

[0028] Furthermore, the step S3 comprises:

[0029] Step S31, constructing the RCCC-WBM model based on the water balance formula. The RCCC-WBM model has the advantages of simple structure, few parameters, easy calibration, and a wide range of applicable climate conditions. For smaller-scale basins, the model can simulate the lumped hydrological process based on the average meteorological elements of the basin; for large-scale basins, the model can grid the study basin and perform distributed simulation of the hydrological process.

[0030] The input data of the RCCC-WBM model are monthly precipitation, monthly temperature, and monthly evapotranspiration. The monthly precipitation is divided into two parts: rainfall and snowfall, which are used to calculate surface runoff and snowmelt runoff. H ), down(T L ) The two critical temperature values ​​are 4℃ and -4℃. When the monthly temperature is between 4℃ and -4℃, the monthly snowfall can be calculated by linear interpolation:

[0031]

[0032] P Ri =P i -P sNi (5)

[0033] Where: P Ri , P SNi , T i are the rainfall, snowfall and average temperature in month i respectively; P i is the potential evapotranspiration in month i.

[0034] Surface runoff and soil moisture are proportional to rainfall. Assuming that groundwater runoff has a time lag of one unit, a linear reservoir simulation is used for calculation. In snowmelt runoff calculation, the snowmelt rate is not only an exponential function of temperature, but also proportional to the amount of snow. Each water source is calculated using equations (6) to (9):

[0035]

[0036] Q gi =k g S i-1 (7)

[0037]

[0038] SNi=SN i-1 +P SNi (9)

[0039] Where: Q Si , Q gi , Q SNi are surface runoff, groundwater runoff and snowmelt runoff in month i (mm); SNi, SN i-1 are the accumulated snow cover in month i and month i-1 (mm); S i-1 is the soil moisture content in month i-1 (mm); S max is the soil water holding capacity of the basin (mm), is S i-1 The upper limit value of k s , k g , k sn They represent the confluence coefficients of surface runoff, groundrunoff and snowmelt runoff respectively.

[0040] One layer of evapotranspiration model can meet the simulation accuracy of actual evapotranspiration. The calculation formula is as follows:

[0041]

[0042] Where: E i is the actual evapotranspiration in month i; E 601 For E 601 Evapotranspiration observed by evapotranspiration dish.

[0043] The total runoff in month i is the sum of surface, underground and snowmelt runoff. The soil moisture content S at the end of month i is i It can be calculated based on the water balance:

[0044] Q ai =Q si +Q gi +Q SNi (11)

[0045] S i =S i-1 +P i +Q ai -E i (12)

[0046] Where: Q ai is the total runoff in month i.

[0047] Step S32: Model parameter calibration and verification. The Nash-Sutcliffe efficiency coefficient (NSE) and relative error (R e ) to reflect the quality of RCCC-WBM simulation results. The closer NSE is to 1, the better R e The closer it is to 0, the better the simulation effect. The calculation formula is:

[0048]

[0049] Where: q ri ,q si , are the measured runoff in month i, the simulated runoff in month i and the average of the measured runoff in the simulation period; n is the series length.

[0050] Step S33, analyze the sensitivity of model parameters. The sensitivity of the four parameters of the RCCC-WBM model is compared by using the local sensitivity analysis method. After determining the optimal parameter value of the model based on the measured hydrological and meteorological data, change the single parameter value within the parameter value range, that is, increase or decrease the parameter to be investigated by an appropriate amount, and then perform hydrological simulation calculations to evaluate the impact of parameter changes on simulation results and objective function changes, and finally determine their sensitivity. The specific calculation is expressed as:

[0051] δ i =F(x i +Δx i )-F(x i ) (i=1,2,3,4) (15)

[0052] Where: x i is the rated value of the i-th parameter; Δx i is the change of the ith parameter; the F function represents the calculated value of the objective function (which can be the Nash-Sutcliffe efficiency coefficient or relative error) after the hydrological model simulation under different parameters; δ i That is, the change value of the objective function, which reflects the sensitivity of the parameters.

[0053] Furthermore, the step S4 comprises:

[0054] Step S41, classification of future climate models. After completing the establishment of the RCCC-WBM model corresponding to step S3, the future climate monthly data processed in step S2, including future monthly precipitation, evapotranspiration, and temperature data, are classified into four climate prediction scenarios.

[0055] Step S42, simulating the future runoff process. The future climate data under different prediction scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5) are respectively input into the model to obtain the future runoff process under different prediction scenarios.

[0056] Furthermore, the step S5 comprises:

[0057] Step S51, analyze the annual distribution characteristics of future runoff. The monthly runoff simulated in step S4 is divided into different periods in the year, and the seasonal index is used to explore its annual distribution characteristics. Due to the seasonal fluctuations of the climate, the runoff changes with the change of seasons, which determines the unevenness of its annual distribution to a certain extent. Here, the annual distribution unevenness coefficient is selected to measure the unevenness of the annual distribution of runoff, and its calculation formula is as follows:

[0058]

[0059] Where: C v is the non-uniformity coefficient, σ is the standard deviation, x i is the measured monthly runoff in month i, billion m 3 ; is the average monthly runoff in a year, billion m 3 ; n is the number of months in the study year, n=12.

[0060] Step 52, analyze the interannual linear trend characteristics of future runoff. The linear trend change analysis of the monthly runoff simulated in step S4 is performed using the Mann-Kendall trend test method. Mann-Kendall is a non-parametric rank correlation statistical test method used to test whether there is a trend in the time series. The specific method and principle are as follows:

[0061] For the hydrological series x, first determine all dual values ​​(x i , x j ; j>i, i=l, 2,..., nl; j=i+l, i+2,..., n) in x i <x j The number of occurrences P constructs the statistic of the Mann-Kendal rank correlation test:

[0062]

[0063] In the formula, n is the number of sequence samples.

[0064] Where: U is the standard normal distribution statistic, τ is the number of positive differences minus the number of negative differences, n is the number of hydrological sequences, and P is the number of x in the hydrological sequence. i <x j The number of occurrences, V ar (τ) is the variance.

[0065] Step 53, analyzing the interannual nonlinear trend characteristics of future runoff. The nonlinear trend test method is used to perform local non-monotonic trend change analysis on the monthly runoff simulated in step S4. The steps of this method are as follows:

[0066] Step 531, suppose that time series A consists of n data (a1, a2, ..., a n ), in which case n is required to be an even number, and the time series is divided into two subsequences A1 and A2 of equal length, where A1 = (a1, a2, …a n / 2), A2=(a 1+ n / 2, a2+n / 2, …, an);

[0067] Step 532, sort each subsequence in ascending order, and calculate the difference between two subsequences to obtain a difference sequence D=A2-A1;

[0068] Step 533, draw a scatter plot of A1, A2 and D in a rectangular coordinate system with the data sequence number (1, 2, 3, ..., n / 2) as the horizontal axis. If all the differences fall on the horizontal axis or deviate very little from the horizontal axis, there is no significant trend in the time series. If the differences are obviously above or below the horizontal axis, the time series has a significant upward or downward trend;

[0069] Step 534, using Fisher's optimal segmentation principle to divide the mean of A1 and A2 sequences into high value area, middle value area, and low value area. At this time, the number of Fisher's optimal segmentation points is 2, which are respectively recorded as Demarcation point1# and Demarcationpoint2#. The area to the left of Demarcation point1# is regarded as the local low value area of ​​the time variable, and the area to the right of Demarcation point2# is regarded as the local high value area of ​​the time variable. The local middle value area of ​​the time variable is the area between Demarcation point1# and Demarcation point2#.

[0070] Step 535, using K L , K M and K H The three index parameters reflect the significance of the non-monotonic trend in the local low value area, the local medium value area and the local high value area. Taking z as an example, if K Z If it is greater than 0, it proves that the local low value area of ​​the time variable is increasing, otherwise it is decreasing. Z The larger the absolute value of , the stronger the significance of the local trend.

[0071] The calculation formula of significance index parameter is:

[0072]

[0073] In the formula, is the average value of the difference in the local low value area; is the average value of the difference in the local median area; is the average value of the difference in the local high value area; is the average value of subsequence A1 in the low value area; is the average value of subsequence A1 in the median area; is the average value of subsequence A1 in the high value area.

[0074] Beneficial effects of the present invention:

[0075] The present invention is based on monthly precipitation data, monthly temperature data, and monthly evapotranspiration data, and combines the RCCC-WBM model with future climate models under different scenarios, which fully solves the problems of the hydrological model used in the existing basin runoff simulation method, such as complex structure, large number of parameters, difficulty in calibration, insufficient resolution, low accuracy, and different regional adaptability; the present invention combines linear trend and nonlinear trend analysis methods to analyze future runoff trends, which solves the problem that the existing runoff change prediction method is not comprehensive in analyzing the time change characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 It is the schematic diagram of the RCCC-WBM model structure;

[0077] Figure 2 Schematic diagram of the accuracy evaluation results of the RCCC-WBM model implemented in the present invention, wherein a is the measured and simulated monthly flow process of the Cuntan Station in the Yangtze River Basin from 1995 to 2004 (rate period), b is the measured and simulated monthly flow process of the Cuntan Station in the Yangtze River Basin from 2005 to 2014, c is the measured and simulated monthly flow process of the Yichang Station in the Yangtze River Basin from 1995 to 2004 (rate period), d is the measured and simulated monthly flow process of the Yichang Station in the Yangtze River Basin from 2005 to 2014, e is the measured and simulated monthly flow process of the Datong Station in the Yangtze River Basin from 1995 to 2004 (rate period), and f is the measured and simulated monthly flow process of the Datong Station in the Yangtze River Basin from 1995 to 2014;

[0078] Figure 3 A line diagram of the future runoff process according to an embodiment of the present invention;

[0079] Figure 4 It is the annual distribution map of future runoff in the embodiment of the present invention;

[0080] Figure 5 The linear change trend characteristic results of future runoff under various scenarios of the embodiments of the present invention;

[0081] Figure 6 These are the nonlinear change trend characteristic results of future runoff under various scenarios of the embodiments of the present invention. DETAILED DESCRIPTION

[0082] The technical method of the present invention is further described below with reference to the accompanying drawings and specific examples.

[0083] The method is illustrated by taking the runoff simulation of the Yangtze River Basin under climate change as an example.

[0084] Step S1. The monthly average precipitation, monthly evapotranspiration, and monthly average temperature data of 684 meteorological stations in the Yangtze River Basin since their establishment were collected, as well as the measured runoff data of three hydrological stations located on the mainstream of the Yangtze River (Cuntan Station, Yichang Station, and Datong Station) from 1995 to 2014. The future climate grid data of China from 2015 to 2100 was collected, including monthly precipitation, maximum temperature, minimum temperature, and average temperature data. After collecting the measured data and future data, the data is first preprocessed.

[0085] Basic information of Yangtze River Basin Control Station:

[0086] Cuntan Station: Cuntan Station is a water control station after the confluence of Jinsha River, Minjiang River, Tuojiang River and Jialing River. It integrates four major functions: flood control station in the two heavy rain areas of western Sichuan and Daba Mountains in the upper reaches of the Yangtze River, flood control station in the main urban area of ​​Chongqing, water and sediment control station for the Three Gorges Project of the Yangtze River, and water environment and water ecology monitoring and control station in the upper reaches of the Yangtze River.

[0087] Yichang Station: Yichang Station was established in March 1946. It is the first major river control station on the mainstream of the middle reaches of the Yangtze River. It is the representative design station for the Three Gorges Project and the Gezhouba Project, which are both major national projects. It is also the general outbound control station for the Three Gorges Water Conservancy Project on the Yangtze River.

[0088] Datong Station: Datong Hydrological Station is the upper boundary of the tidal section of the Yangtze River estuary and the control station before the Yangtze River enters the sea. The runoff is composed of water from the upstream Yichang Station, tributaries such as the Han River, Dongting Lake and Poyang Lake, and other small tributaries and the confluence of the Yichang-Datong Station area, representing the amount of water entering the sea in the Yangtze River Basin. The runoff data of each station are calculated as monthly runoff for unification.

[0089] Step S2. The preprocessed future climate grid data of China from 2015 to 2100 are clipped using ArcGIS software to obtain the future climate grid data of the Yangtze River Basin from 2015 to 2100, and then all variables in the 19 climate models are interpolated to a regular grid of 0.1°×0.1° using bilinear interpolation to make it consistent with the spatial resolution of the meteorological driving data set. The Delta downscaling method is then used to downscale the climate data of the 19 climate models in CMIP6 (as shown in Table 1) under four projection scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5), and then the average of the meteorological data of the 19 climate models that have been processed is calculated, and the multi-model average monthly precipitation, maximum temperature, minimum temperature, and average temperature data are combined with the Hargreaves formula to obtain the multi-model average monthly evapotranspiration data.

[0090] Table 1 Basic information of climate model

[0091]

[0092]

[0093] Step S3. Based on various types of measured data in the Yangtze River Basin after pretreatment, the RCCC-WBM model is constructed, and the measured runoff data of Cuntan Station, Yichang Station, and Datong Station from 1995 to 2014 are used to calibrate and verify the model. The calibration period is usually about ten years, and there will be missing values ​​in the middle of the long sequence obtained. In order to ensure the objectivity of the data, the continuous part is selected. The calibration period of the RCCC-WBM model constructed by the present invention is set to January 1, 1995 to December 31, 2004; the verification period is set to January 1, 2005 to December 31, 2014, and the simulation step length is monthly.

[0094] The data processed in step S2, i.e., monthly precipitation data for many years, monthly potential evapotranspiration, monthly average temperature, and monthly runoff, are input into the table in the corresponding format, and the table name is changed to the corresponding format and placed in the corresponding folder, and the corresponding data path is changed, the table range is selected, and the loading and writing of meteorological data are completed.

[0095] Figure 2 The Yangtze River Basin is divided into the upper, middle and lower reaches with Cuntan, Yichang and Datong stations as control stations, and the accuracy evaluation results of the established RCCC-WBM model in various regions of the Yangtze River Basin are shown. It can be seen from the figure that the RCCC-WBM model based on water balance constructed by the present invention has high accuracy and can well simulate the runoff situation in different regions of the Yangtze River Basin. The efficiency coefficient of the Nash-Sutcliffe model at the rate period exceeds 85%, and the average relative error is also controlled to be small, as shown in Table 2, indicating that the model has good regional adaptability in the Yangtze River Basin.

[0096] Table 2 Simulation results of each basin

[0097]

[0098]

[0099] Step S4. Based on the completed RCCC-WBM model with high simulation accuracy for the Yangtze River Basin, the Datong Station is used as the control station for the entire Yangtze River Basin. The processed future climate data are classified according to four climate projection scenarios (SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5), representing low, medium and high emission scenarios respectively, and are respectively introduced into the RCCC-WBM model to simulate the future runoff of the Yangtze River Basin under climate change conditions. Finally, the monthly runoff series of the Yangtze River Basin from 2015 to 2100 under different projection scenarios are obtained, see Figure 3 .

[0100] The CMIP6 model adopted in the present invention not only has an unprecedentedly large number of global climate models (GCMs) participating in the CMIP6 plan compared to the previous generation model plan (CMIP5), but each GCMs also represents the forefront of the development of current climate system models, and the model design scientific experiments are also the most complete, including more physical processes and feedback mechanisms, combining shared socioeconomic pathways (SSPs) with representative concentration pathways (RCPs), and constructing 7 new SSPs scenarios (SSP1-1.9, SSP1-2.6, SSP4-3.4, SSP2-4.5, SSP4-6.0, SSP3-7.0, SSP5-8.5), which also greatly improves the simulation ability of extreme weather and climate events, and the prediction results are more reliable. The present invention selects four climate projection scenarios: SSP1-2.6 (SSP1+RCP2.6), SSP2-4.5 (SSP2+RCP4.5), SSP3-7.0 (SSP3+RCP7.0) and SSP5-8.5 (SSP5+RCP8.5), which represent low, medium and high emission scenarios respectively. The predicted runoff scale is fine, which is the monthly scale, can well reflect the distribution characteristics of runoff within the year, and is more practical.

[0101] Step S5. After obtaining the runoff series of the Yangtze River Basin in the future period under different prediction scenarios, the seasonal index is first used to analyze the annual distribution of future runoff, such as Figure 3 The results show that there is strong seasonality in all scenarios. In the SSP1-2.6, SSP2-4.5, and SSP5-8.5 scenarios, the seasonal index shows a downward trend, indicating that the runoff is more evenly distributed throughout the year. There is no obvious trend in the SSP3-7.0 scenario. Linear trend analysis and nonlinear trend analysis were then used to analyze the temporal evolution of future runoff in the Yangtze River Basin. Under the linear trend test, the annual runoff of the Yangtze River showed an upward trend in all scenarios; under the nonlinear trend test, the difference sequence of the annual runoff of the Yangtze River in all scenarios was absolutely above the horizontal line y=0, such as Figure 4-5 The annual runoff of the Yangtze River generally shows an upward trend, and the low-value area, medium-value area and high-value area of ​​annual runoff all show an upward trend. Figure 6 .

[0102] The above description is only an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for simulating the temporal and spatial variation of basin runoff under climate change based on the RCCC-WBM model, characterized in that: The following steps are involved: S1. Obtain the measured precipitation, runoff, temperature, and evapotranspiration data of the study area, including monthly average precipitation, monthly runoff, monthly average temperature, monthly average evapotranspiration, watershed area, and future climate data; future climate data include future monthly precipitation, maximum temperature, minimum temperature, and average temperature, and pre-process the data; S2. Perform clipping, interpolation, downscaling and multi-model averaging on the corresponding future climate data in step S1, and calculate the corresponding future evapotranspiration data using the Hargreaves formula; S3. Based on the measured precipitation, runoff, temperature and evapotranspiration data corresponding to step S1, the RCCC-WBM model based on water balance is constructed, and the measured monthly runoff data of the corresponding stations in the study area are used to calibrate and verify the model parameters; S4. Classify the future climate monthly data processed in step S2 into four climate prediction scenarios, input them into the RCCC-WBM model established in step S3, and simulate the future runoff; S5. Use seasonal index, Mann-Kendall linear trend and nonlinear trend test methods to analyze the temporal changes of the monthly future runoff series obtained in step S4.

2. The method for simulating the temporal and spatial variation of watershed runoff under climate change based on the RCCC-WBM model according to claim 1, characterized in that: The preprocessing of meteorological data in step S1 includes: Step 11, perform missing value test, consistency test and white noise processing on the measured precipitation, temperature and evapotranspiration data; Step 12, interpolate and extend the measured precipitation, temperature, and evapotranspiration data, and convert the units of the measured monthly runoff data, converting the flow units into runoff depth units.

3. The method for simulating the spatiotemporal variation of watershed runoff under climate change based on the RCCC-WBM model according to claim 1, characterized in that: The step S2 comprises: Step S21, cutting the acquired future climate data to the corresponding research area; Step S22, using bilinear interpolation to interpolate all variables in the 19 climate models in CMIP6 into a grid so that the spatial resolution is consistent with the meteorological driving dataset; Step S23, using the Delta downscaling method to perform downscaling analysis on the climate data of 19 climate models in CMIP6 under four projection scenarios, namely SSP1-2.6, SSP2-4.5, SSP3-7.0 and SSP5-8.5; Step 24, averaging the future climate data after the downscaling analysis according to multiple models to obtain multi-model average future climate data; Step S25, calculating future evapotranspiration data, calculating the future evapotranspiration data by using the Hargreaves formula to calculate the multi-model average future climate data.

4. The method for simulating the temporal and spatial variation of watershed runoff under climate change based on the RCCC-WBM model according to claim 1, characterized in that: The step S3 comprises: Step S31, constructing a RCCC-WBM model based on a water balance formula; Step S32, model parameter calibration and verification; Step S33, analyzing the sensitivity of model parameters.

5. The method for simulating the temporal and spatial variation of watershed runoff under climate change based on the RCCC-WBM model according to claim 3 is characterized in that: The step S4 comprises: Step S41, classification of future climate models, after completing the establishment of the RCCC-WBM model corresponding to step S3, the future climate monthly data processed in step S2, including the future monthly precipitation, evapotranspiration, and temperature data, are classified into four climate prediction scenarios; Step S42, simulating the future runoff process, inputting the future climate data under different prediction scenarios into the model respectively, and obtaining the future runoff process under different prediction scenarios.

6. The method for simulating the temporal and spatial variation of watershed runoff under climate change based on the RCCC-WBM model according to claim 5, characterized in that: The step S5 comprises: Step S51, dividing the monthly runoff simulated in step S4 into different periods within a year, and using seasonal index to explore its distribution characteristics within the year; Step 52, using the Mann-Kendall trend test method to perform a linear trend change analysis on the monthly runoff simulated in step S4; Step 53, using a nonlinear trend test method to perform a local non-monotonic trend change analysis on the monthly runoff simulated in step S4.

Images