Indigenous fish species distribution simulation method coupled with hydrology and water temperature

Abstract

The invention discloses an indigenous fish species distribution simulation method coupled with hydrology and water temperature, and relates to the field of fish species distribution models.The indigenous fish species distribution simulation method comprises the steps that a watershed distributed hydrothermal coupling model of a target area is constructed, and hydrothermal environment variables of the target area are determined based on the watershed distributed hydrothermal coupling model; determining and screening predictive variables related to the target indigenous fishes, wherein the predictive variables comprise hydrothermal environment variables; point location data of target indigenous fishes in a target area are collected, and a target indigenous fish relative abundance distribution model is constructed in combination with the screened predictive variables; and simulating and predicting the relative abundance of the future target indigenous fish by using the relative abundance distribution model. According to the method, flow and water temperature habitat elements directly related to survival of the indigenous fishes can be comprehensively considered, the problem of simulation and prediction of relative abundance of the indigenous fishes under climate change is solved, and a basis is provided for habitat protection of the indigenous fishes under climate change.

Description

Technical Field

[0001] The present invention relates to the technical field of fish species distribution models, and more particularly to a method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature. Background Art

[0002] Flow and water temperature are key habitat factors affecting fish. Reduced flow affects the dispersal of fish eggs, which in turn affects the composition, diversity, spawning, and size structure of fish populations. Increased winter flow causes fish eggs to be washed away by riverbed gravel, reducing the survival rate of eggs of autumn-spawning fish. Increased seasonal flow triggers spawning migrations, increasing access to feeding and spawning habitats. Fish are cold-blooded animals, and their physiology is inextricably linked to the temperature of their surroundings. Temperature conditions regulate various aspects of fish physiology, such as growth, reproduction, and activity. Changes in water temperature can affect fish throughout their life cycle.

[0003] However, under the backdrop of global climate change, the hydrological and water temperature regimes of rivers are undergoing significant changes. Rising temperatures intensify water-air exchange, leading to higher water temperatures. Natural processes such as glacier melting, changes in snow cover, and increased precipitation can affect changes in runoff. Altered flow and water temperature patterns under climate change will affect fish migration, reproduction, growth, and distribution in various ways. However, traditional species distribution models (SDMs) predict the distribution of suitable fish habitats in geographic space by analyzing the relationship between habitat distribution and macroclimatic factors (such as precipitation and temperature), but they ignore the flow and water temperature habitat factors that are directly related to fish survival. Furthermore, existing species distribution models use binary "presence / absence" data to predict distribution ranges, making it difficult to reflect changes in fish relative abundance and the gradient characteristics of suitable fish habitats.

[0004] Therefore, how to simulate and predict the relative abundance of target indigenous fish species by taking into account flow and water temperature habitat factors under climate change conditions is an urgent problem that technicians in this field need to solve. Summary of the Invention

[0005] In view of this, the present invention provides a method for simulating the distribution of indigenous fish species by coupling hydrology and water temperature. Based on the relationship between hydrothermal and other environmental factors and the relative abundance of fish populations, it quantitatively describes the species distribution of target fish under climate change, and provides a scientific basis for the identification of key fish habitats, ecological flow regulation and maintenance of river biodiversity security under climate change.

[0006] In order to achieve the above object, the present invention provides the following technical solutions:

[0007] The present invention discloses a method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature, and the specific steps are as follows:

[0008] Constructing a watershed-distributed hydrothermal coupling model for the target area, and determining the hydrothermal environmental variables of the target area based on the watershed-distributed hydrothermal coupling model;

[0009] determining and screening predictor variables related to target indigenous fish, wherein the predictor variables include the hydrothermal environment variables;

[0010] The point data of target indigenous fish in the target area are collected, and a relative abundance distribution model of the target indigenous fish is constructed by combining the screened prediction variables; the relative abundance distribution model is used to simulate and predict the relative abundance of the target indigenous fish in the future.

[0011] Furthermore, the construction of a distributed water-heat coupling model for the target area specifically includes:

[0012] Construct a spatial database based on the digital elevation data, land use type data, soil type data and glacier vector data of the target area;

[0013] Construct an attribute database based on soil attribute data, meteorological data, runoff and water temperature observation data of the target area;

[0014] Establishing a SWAT model based on the spatial database and the attribute database;

[0015] The hydrological module and the water temperature module in the SWAT model are improved to obtain the basin distributed water-heat coupling model.

[0016] Furthermore, the improvement of the hydrological module is as follows: adding a glacier module to the hydrological module;

[0017] The water balance equation of the hydrological module is as follows:

[0018]

[0019] Among them, SW t is the soil moisture content at the end of day t; SW0 is the initial soil moisture content; Pr i , Q surf,i 、E i , I i , Q lat,i and Q gr,i are the precipitation, surface runoff, evapotranspiration, soil infiltration, intersoil flow and base flow on day i, all in mm;

[0020] The glacier melting equation of the glacier module is as follows:

[0021] W g '-W g = -(1-f)M+FS;

[0022] Among them, W g is the initial water equivalent, W g ' is the updated water equivalent, M is the glacier melt volume, f is the re-solidification ratio of glacier meltwater, f=0.2, F is the accumulated amount of glacier material, and S is the sublimation rate of the glacier; the glacier melt volume M is equivalent to precipitation and superimposed on the precipitation in the water balance equation, thereby adding the simulation of glacier runoff to the water cycle simulation and obtaining an improved hydrological module.

[0023] Furthermore, the water temperature module is improved by replacing the temperature calculation formula of the original water temperature module with a water temperature calculation formula that considers the influence of different runoff components on water temperature. The formula is:

[0024]

[0025] Among them, T water is the water temperature, j represents the sub-basin number; T air is the average daily temperature; KK is the overall heat transfer coefficient, slr is the solar radiation; TT is the time it takes for water to pass through the sub-basin; η is an additional parameter for the water-air heat transfer process; ρ w is the density of water, is the specific heat capacity of water, H is the water depth; Twater intial is the temperature of the upstream sub-basin.

[0026] Furthermore, the calculation formula for the upstream sub-basin temperature is:

[0027]

[0028] Among them, T waterupstrem is the water temperature of the river entering the sub-basin; Q outlet is the outlet flow of the sub-basin; sub_wyld is the total water production of the sub-basin; T w,local is the local water temperature of the sub-basin, and the calculation formula is:

[0029]

[0030] Among them, T w,local represents the local water temperature of the sub-basin; sub_snow is the snowmelt contribution of the sub-basin, sub_gla is the glacier melt contribution of the sub-basin, sub_gw is the groundwater contribution of the sub-basin, sub_surq is the surface runoff contribution of the sub-basin, and sub_latq is the lateral soil flow of the sub-basin; α, β, λ, and κ are the calibration coefficients of the relative contribution of each runoff component to the local water temperature; T sm represents the snowmelt temperature, T gw represents the groundwater temperature, T gla represents the glacier melt runoff temperature, T air,lag represents the lagged daily average temperature, and the calculation formula is:

[0031] T air,lag,i =d×T air,lag(i-1) +(1-d)T air,i ;

[0032] Where d is the hysteresis coefficient, T air is the average daily temperature, and i represents the number of simulation days.

[0033] Furthermore, the determination and screening of predictor variables related to target indigenous fish species specifically include:

[0034] Determine the bioclimatic variables, human activity variables, physical geographical variables and hydrothermal environment variables of the target area;

[0035] Then, Spearman correlation coefficient analysis was performed on each variable, and environmental variables with correlation coefficient r ≥ |0.8| were eliminated;

[0036] Variance inflation factors of the remaining variables were calculated, and predictor variables with variance inflation factors of a higher order of magnitude than other variables were eliminated in each round. The remaining variables were used to model the target indigenous fish species.

[0037] Furthermore, the construction of the relative abundance distribution model of the target indigenous fish species specifically includes:

[0038] Collect point data of target indigenous fish in the target area and perform sparse processing;

[0039] Based on the sparsely processed point data and the screened predictor variables, the maximum entropy model was used to predict the suitable habitat of the target indigenous fish.

[0040] Based on the sparsely processed point data and the screened predictor variables, a generalized linear mixed model was used to simulate the abundance distribution of target indigenous fish.

[0041] Furthermore, the regulation ratio and feature combination of the maximum entropy model are adjusted by the AIC information criterion, and the regulation ratio value and feature combination parameters with the smallest AIC value are selected to construct the final model.

[0042] Furthermore, the formula of the generalized linear mixed model is:

[0043] y=Xτ+Zu+e;

[0044] Among them, y is the vector of observed values ​​of fish relative abundance, τ is the vector of all fixed effects, X is the correlation matrix of fixed effects, u is the vector of all random effects, Z is the correlation matrix of random effects, and e is the random residual vector.

[0045] It can be seen from the above technical solution that compared with the existing technology, the present invention discloses a method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature, which can comprehensively consider the flow and water temperature habitat factors directly related to the survival of indigenous fish, solve the problem of simulating and predicting the relative abundance of indigenous fish under climate change, and provide a basis for the protection of indigenous fish habitats under climate change. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0047] Figure 1 Schematic diagram of the overall process of an embodiment of the present invention.

[0048] Figure 2 Schematic diagram of relative abundance simulation according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0050] The embodiment of the present invention discloses a method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature. Figure 1 The specific steps are as follows:

[0051] Construct a watershed-distributed hydrothermal coupling model for the target area, and determine the hydrothermal environmental variables of the target area based on the watershed-distributed hydrothermal coupling model;

[0052] Identify and screen predictor variables related to target indigenous fish species, including hydrothermal environment variables;

[0053] Collect point data of target indigenous fish in the target area, combine it with the screened prediction variables, and construct a relative abundance distribution model of target indigenous fish; use the relative abundance distribution model to simulate and predict the relative abundance of target indigenous fish in the future.

[0054] In a specific embodiment, constructing a distributed water-heat coupling model for a target area includes:

[0055] Construct a spatial database based on the digital elevation data, land use type data, soil type data and glacier vector data of the target area;

[0056] Construct an attribute database based on soil attribute data, meteorological data, runoff and water temperature observation data of the target area;

[0057] Establish SWAT model based on spatial database and attribute database;

[0058] The hydrological module and water temperature module in the SWAT model were improved to obtain a basin-distributed water-heat coupling model.

[0059] Specifically, the construction of the spatial database includes digital elevation data, land use type data, soil type data and glacier vector data. Among them, the digital elevation data comes from the ASTER GDEMV2 digital elevation data (spatial resolution 30m) provided by the Geospatial Data Cloud Platform of the Computer Network Information Center of the Chinese Academy of Sciences. ArcGIS10.2 is used for image fusion, splicing, cropping, filling, projection and other processing to obtain the digital elevation data of the watershed range; the land use type data comes from the national land use type remote sensing monitoring spatial distribution data (spatial resolution 30m) provided by the Resource and Environmental Science and Data Center of the Chinese Academy of Sciences, including 6 first-level types of cultivated land, forest land, grassland, water area, construction land and unused land, and 21 second-level types. ArcGIS10.2 is used for image cropping, projection, reclassification and other processing to obtain the watershed range in line with SWA Land use data conforming to the SWAT model land use classification standard were used. Soil type data were based on the China Soil Dataset (V1.1) (2009) from the World Soil Database. The data source within China was the 1:1,000,000 soil data (spatial resolution 1000 m) provided by the Second National Land Survey in Nanjing. ArcGIS 10.2 was used for image cropping, projection, attribute association, and reclassification to obtain basin-wide soil type data that conformed to the SWAT soil classification standard. The glacier dataset was based on the First Glacier Inventory of China (GIA-I) and the Second Glacier Inventory of China (GIA-II). GIA-I refers to glacier conditions from the 1960s and 1970s, while GIA-II refers to glacier conditions from 2007 to 2009. A glacier dataset from the GIA-I era was compiled to initialize the simulation, and glaciers from the GIA-II era were used to validate the simulation. Because the glacier types in the original soil and land use data differ significantly from the glacier area and distribution in the glacier inventory data, accurate simulation of basin runoff using the SWAT model was required. The first and second glacier vector data were used to replace the glacier snow cover type (ICE) in the soil data. The first and second glacier vector data were used to replace the glacier type data (CNJX) in the land use type.

[0060] The attribute database constructed includes soil attribute data, meteorological data, and runoff and water temperature observation data. Soil attribute data, including the number of soil layers, soil texture, maximum root depth in the soil profile, soil particle composition, anion exchange porosity, depth from the surface to the base layer, and organic carbon content, were obtained from the China Soil Dataset (V1.1) of the World Soil Database. Soil bulk density, effective water holding capacity, and saturated hydraulic conductivity were calculated using SPAW software. Meteorological data were obtained from the China Meteorological Data Network's China Surface Climate Data Daily Dataset. Daily data on rainfall, relative humidity, temperature, and wind speed were processed into txt files, and index files corresponding to each file were created for easy reference during model writing. Runoff and water temperature observation data, including daily average runoff and daily average water temperature, were obtained from the Hydrological Yearbook. These data were used to calibrate the SWAT model parameters and verify the model simulation accuracy.

[0061] The SWAT model is a distributed soil and water assessment model based on physical mechanisms. The main modules of the SWAT model include hydrological module, sediment transport, pollutant transport and agricultural management measures. The hydrological module is designed based on the water balance framework. The river network is defined based on the threshold method, and the minimum catchment area threshold for sub-basin division is set. Outlet points are deleted and added according to actual conditions. Each sub-basin is divided into multiple hydrological response units. By setting thresholds for the proportion of land use type area and soil type area to the sub-basin area, types below the set area threshold are discarded, and types larger than the threshold are superimposed according to spatial distribution to form multiple hydrological response units.

[0062] In a specific embodiment, the improvement of the hydrological module is as follows: a glacier module is added to the hydrological module; specifically, a glacier runoff simulation is added to the water cycle simulation to improve the hydrological module, Visio Studio 2019 is used to recompile the SWAT source code, a glacier module is constructed based on the glacier ablation equation, and it is written into the SWAT source code file based on the glacier mass balance, and the original "swat.exe" is replaced with the newly generated "swat.exe", and the glacier runoff simulation is considered to improve the hydrological module.

[0063] The water balance equation of the hydrological module is as follows:

[0064]

[0065] Among them, SW t is the soil moisture content at the end of day t; SW0 is the initial soil moisture content; Pr i , Q surf,i 、E i , I i , Q lat,i and Q gr,iare the precipitation, surface runoff, evapotranspiration, soil infiltration, intersoil flow and base flow on day i, all in mm;

[0066] The glacier melt equation of the glacier module is as follows:

[0067] W g '-W g = -(1-f)M+FS;

[0068] Among them, W g is the initial water equivalent, W g ' is the updated water equivalent, M is the glacier melt volume, f is the re-solidification ratio of glacier meltwater, f = 0.2, F is the accumulated glacier material, and S is the sublimation rate of the glacier; the glacier melt volume M is equivalent to precipitation and superimposed on the precipitation in the water balance equation, thereby adding the simulation of glacier runoff to the water cycle simulation and obtaining an improved hydrological module.

[0069] In a specific embodiment, the original water temperature module formula is: T water =5.0+0.75T air The original water temperature module is based on the linear relationship between air temperature and water temperature, ignoring the influence of different runoff components on it. Therefore, the original water temperature module formula is improved. The improvement of the water temperature module is as follows: the temperature calculation formula of the original water temperature module is replaced with a water temperature calculation formula that considers the influence of different runoff components on water temperature. The formula is:

[0070]

[0071] Among them, T water is the water temperature (unit: °C), j represents the sub-basin number; T air is the average daily temperature (unit: °C); KK is the overall heat transfer coefficient, which needs to be calibrated and ranges from 0 to 1. The value of KK depends on the relationship between the water flow and the air temperature. For example, if the water flow temperature is roughly the same as the air temperature, KK is 1; slr is the solar radiation (unit: MJ / m 2 ); TT is the time it takes for water to pass through the sub-basin (unit: h); η is an additional parameter of the water-air heat transfer process; ρ w is the density of water (unit: kg / m 3 ), is the specific heat capacity of water (unit: J / kg℃), H is the water depth; Twater intial is the temperature of the upstream sub-basin. air and the influence of hydrological contribution make the final T water When the temperature is less than 0℃, the river temperature model will water Set to 0.1°C. T water is also assumed to be the temperature of water flowing downstream to the sub-basin and further along the river network.

[0072] The new formula was written into the SWAT source code file with glacial runoff, and the original "swat.exe" was replaced with the newly generated "swat.exe". Based on the constructed hydrological model, the changes in different runoff components were simulated, and a water temperature module was constructed according to the formula. At the same time, the SWAT-CUP software was used to calibrate and verify the water temperature module parameters, thereby simulating the water temperature changes in different sub-basins.

[0073] In a specific embodiment, the calculation formula of the upstream sub-basin temperature is:

[0074]

[0075] Among them, T waterupstrem is the water temperature of the river entering the sub-basin; Q outlet is the outlet flow of the sub-basin; sub_wyld is the total water yield of the sub-basin, obtained from the hydrological module of the SWAT model; T w,local is the local water temperature of the sub-basin, and the calculation formula is:

[0076]

[0077] Among them, T w,local represents the local water temperature of the sub-basin; sub_snow is the snowmelt contribution of the sub-basin (unit: m 3 / d), sub_gla is the sub-basin glacier melting contribution (unit: m 3 / d), sub_gw is the groundwater contribution of the sub-basin (unit: m 3 / d), sub_surq is the surface runoff contribution of the sub-basin (unit: m 3 / d), sub_latq is the lateral soil flow in the sub-basin (unit: m 3 / d); α, β, λ, and κ are the calibration coefficients of the relative contribution of each runoff component to the local water temperature, which are dimensionless and need to be calibrated; T sm represents the snowmelt temperature, T gw represents the groundwater temperature, T gla represents the temperature of glacier melt runoff. Assuming that the glacier melt temperature is 0.1°C, the groundwater temperature is usually 1-2°C higher than the average annual temperature in a region and can be estimated based on climate input data; T air,lag represents the lagged daily average temperature, and the calculation formula is:

[0078] T air,lag,i =d×T air,lag(i-1) +(1-d)T air,i ;

[0079] Where d is the hysteresis coefficient, T airis the daily mean temperature, and i represents the number of simulation days. Lag (days) is a calibration parameter used to consider the effects of surface runoff delay and soil water inflow into streams. If T air,lag =0℃, set to 0.1℃.

[0080] In a specific embodiment, the effectiveness of the improved SWAT model is evaluated by:

[0081] The Nash coefficient (NSE) and correlation coefficient (R²) were used to evaluate the applicability of the SWAT model. The NSE ranges from -∞ to 1. An NSE closer to 1 indicates a good simulation and high model credibility. An NSE closer to 0.5 indicates that the model simulation results are close to the average level of observed values ​​and that the overall model results are reliable. R² indicates the degree and direction of linear correlation between simulated and measured values. A value closer to 1 indicates a better correlation between simulated and measured values.

[0082] In a specific embodiment, determining and screening predictive variables related to target indigenous fish species specifically includes:

[0083] Determine the bioclimatic variables, human activity variables, physical geographical variables and hydrothermal environment variables of the target area;

[0084] Then, Spearman correlation coefficient analysis was performed on each variable, and environmental variables with correlation coefficient r≥|0.8| were eliminated to avoid autocorrelation and multicollinearity between variables that would lead to overfitting of the model;

[0085] The variance inflation factors of the remaining variables are calculated, and predictor variables with variance inflation factors that are orders of magnitude higher than those of other variables are eliminated in each round. The target indigenous fish species are then modeled using the remaining variables. Furthermore, the present invention eliminates less important variables in each round of modeling. Finally, the response curve of each variable is examined to eliminate variables with unreasonable response mechanisms.

[0086] Specifically, the bioclimatic variables are commonly used in the field of ecological niche modeling from the WorldClim database, including 11 temperature-related variables and 8 precipitation-related variables. The data for the future period adopt the high (SSP585), medium (SSP370), and low (SSP126) emission pathways of the Shared Socioeconomic Pathways (SSPs). To fully represent the overall characteristics of the CMIP6 model ensemble, the average values ​​of the five global climate models UKESM1-0-LL, MRI-ESM2-0, MPI-ESM1-2, IPSL-CM6-LR, and GFDL-ESM4 are integrated to simulate climate change in the two future periods of 2040-2060 (2050s) and 2080-2100 (2090s).

[0087] The human activity variables are population density (PopD), population size (PopC), human impact index (HII), proportion of human-unaffected areas (HIP), river network connectivity coefficient (RNC), cultivated land area (CLA) and development pressure index (DPI).

[0088] The physical geographical variables are land elevation (DEM), river classification (RO), and watershed area (WTA). The river classification (RO) is made based on river classification information, and the river network connectivity coefficient (RNC) is a quantitative coefficient within the dendritic ecological network (vertical) that takes into account the impact of dams.

[0089] The hydrothermal environmental variables are the flow and water temperature data simulated by the distributed hydrothermal coupling model of the basin, and five groups of 66 change indicators of hydrology and water temperature constructed based on the IHA method (hydrological and water temperature change index). The IHA method includes five groups, a total of 33 hydrological indicators and 33 water temperature indicators, which are calculated based on the water temperature and runoff simulated by the above-mentioned distributed hydrothermal coupling model of the basin. The five groups of indicators include: ① Monthly average flow and water temperature, a total of 24, respectively, the median flow and water temperature of each month from January to December; ② Annual extreme flow and water temperature and duration, a total of 24, respectively, the annual minimum flow and water temperature on the 1st, 3rd, 7th, 30th, and 90th day, the annual maximum flow and water temperature on the 1st, 3rd, 7th, 30th, and 90th day, the number of days of dry-up (the number of days with frozen water temperature), the base flow index (the water temperature is the base flow temperature, that is, the water temperature in the dry season (representing the base flow as the main flow) ) and the annual average water temperature, reflecting the difference between groundwater and annual water temperature, and reflecting the temperature regulating effect of groundwater); ③ Annual extreme flow water temperature occurrence time, a total of four, namely, the occurrence time of the annual minimum flow water temperature, and the occurrence time of the annual maximum flow water temperature; ④ The frequency and duration of high and low flow water temperatures, a total of eight, namely, the number of low flow water temperature occurrences per year, the average history of low flow water temperature per year, the number of high flow water temperature occurrences per year, and the average duration of high flow water temperature per year; ⑤ Flow water temperature change rate and frequency, a total of six, including the flow water temperature rise rate, flow water temperature drop rate, and the number of flow water temperature reversals.

[0090] In a specific embodiment, constructing a relative abundance distribution model of target indigenous fish species specifically includes:

[0091] We collected location data for target native fish species in the target area. To avoid model prediction bias caused by multiple distribution points within the same grid, we used the "spThin" package in R to thin the original distribution points, retaining only one distribution point per square kilometer. This reduced the impact of spatial autocorrelation on model predictions. Ultimately, we retained independent distribution points for subsequent species distribution model construction.

[0092] Based on the sparsely processed point data and the screened predictor variables, the maximum entropy model was used to predict the suitable habitat of the target indigenous fish.

[0093] Based on the sparsely processed point data and the screened predictor variables, a generalized linear mixed model was used to simulate the abundance distribution of target indigenous fish.

[0094] In a specific embodiment, the control factor and feature combination of the maximum entropy model are adjusted by the AIC information criterion, and the control factor value and feature combination parameters with the smallest AIC value are selected to construct the final model.

[0095] Specifically, the maximum entropy model MaxEnt3.4.3 (Phillips and Dudík, 2008) was used to predict suitable habitats for target indigenous fish species. After inputting the distribution site and environmental variable layers, the number of replicates was set to 5, and class selection cross-validation was repeated. Data were output in Cloglog format. Jackknife results were also input to assess the contribution of each environmental variable to the distribution of the target indigenous fish species, plot response curves, and generate prediction graphs. The area under the receiver operating characteristic (ROC) curve (AUC) was used to evaluate the quality of the model predictions. The Cloglog threshold of the 10% existence point training set was selected as the basis for dividing suitable habitats and unsuitable habitats. The areas with suitability values ​​lower than the threshold were divided into unsuitable habitats, and vice versa. The other model parameters were the default values. The optimization of the MaxEnt model is closely related to its regulation frequency multiplication (RM) and feature combination (FC) parameters. The RM and FC parameters of the MaxEnt model were adjusted by calling the "ENMeval" package in the R software to test the AIC information criterion of the modified MaxEnt model under different parameter conditions, thereby evaluating the excellence of the model. The AIC information criterion gave priority to the model with the smallest AIC value. In the present invention, RM was set to 0.5-4, increasing by 0.5 each time, and 6 feature combinations (FCs) were used, namely: L (linear features), LQ (linear and quadratic features), H (hinge features), LQH (linear, quadratic and hinge features), LQHP (linear, quadratic, hinge and product features) and LQHPT (linear, quadratic, hinge, product and threshold features). The RM value and FC parameter structure with the smallest AIC value were selected. The final maximum entropy model was constructed. 1 and 0 were used to represent the suitable and unsuitable habitats of the target indigenous fish, respectively. Overlay analysis was performed in ArcGIS 10.2, and the calculation formula was as follows: Y = SHP × 10 + SHL. Among them, Y is the habitat transfer data of the target indigenous fish, SHP and SHL are the suitable habitat data under a certain path in the current period and the future period, respectively. The Y values ​​are 0, 1, 10, and 11, respectively, representing the changes in suitable habitats in different periods: (1) Unsuitable habitat: From now to 2050s, from now to 2090s are always the target (1) The area is an unsuitable habitat for the target indigenous fish species, and the value of Y is 0; (2) New suitable habitat: The area that is currently unsuitable for the target indigenous fish species will be converted into a suitable habitat by the 2050s and 2090s, and the value of Y is 1; (3) Loss of suitable habitat: The area that is currently suitable for the target indigenous fish species will be converted into an unsuitable habitat by the 2050s and 2090s, and the value of Y is 10; (4) Retention of suitable habitat: The area that is currently suitable for the target indigenous fish species in the 2050s and 2090s will always be a suitable habitat for the target indigenous fish species, and the value of Y is 11.

[0096] In a specific embodiment, the formula of the generalized linear mixed model is:

[0097] y=Xτ+Zu+e;

[0098] Where y is the vector of observed fish relative abundance, τ is the vector of all fixed effects (including water temperature, flow, and bioclimatic variables), X is the correlation matrix of fixed effects, u is the vector of all random effects, Z is the correlation matrix of random effects (screened human activity variables + physical geographic variables), and e is the random residual vector.

[0099] Specifically, the remaining human activity and physical geography variables were used as random effects for model fitting. Linear mixed models were constructed based on different combinations of random effects. The F-test (a test for the F-distribution) was used to assess the significance of each factor in the relationship between the environmental factor and the relative abundance of the target native fish. The AIC was used to compare model fit, while the root mean square error (RMSE) was used to determine the optimal model; the closer the AIC value is to 0, the smaller the difference between the model's predicted values ​​and the true values ​​in the sample, indicating a better model fit. All models were constructed using the "lme4" package in R software. Variables such as water temperature and discharge under different future climate change scenarios were standardized to the mean of historical data. The standardized future fixed effects were multiplied by the trained fixed effect coefficients to calculate the predicted relative abundance of the target native fish species. The future relative abundance of the target native fish species was predicted using the "predict" function in the "lme4" package in R software, assuming that the variances of the random effects of human activity and physical geography were consistent with those in the historical model.

[0100] In a specific embodiment, the Bosten Lake Basin is used as the target area, and the Xinjiang naked lip fish is used as the target indigenous fish for illustration.

[0101] The process of constructing the spatial database is described in detail in the above embodiment.

[0102] In the process of constructing the attribute database, soil attribute data and meteorological data are specifically referred to the above embodiment; runoff and water temperature observation data are derived from the hydrological yearbook, and the measured runoff and water temperature data of Dashankou Station in the Bosten Lake Basin are selected, including the daily average runoff and daily average water temperature, for calibrating the parameters of the SWAT model and verifying the model simulation accuracy.

[0103] The SWAT model was established, specifically including: generating hydrological response units by dividing sub-basins, defining the river network based on the threshold method, setting the minimum catchment area threshold for sub-basin division, and setting the threshold to 4000 Ha. According to the actual situation, the outlet points of tributaries were deleted and added, and the upper reaches of Lake Bosten were finally divided into 43 sub-basins; using the "land use / soil / slope definition", land use data and soil data were loaded, and a corresponding relationship was established with the index table; using the "HRU definition", the "Multiple HRUs" division method was selected, and the thresholds of land use, soil and slope were all set to 5%, and finally 445 hydrological response units were generated. The meteorological data set was loaded, and the daily data of rainfall and temperature were processed into txt format documents. At the same time, index files corresponding to each document were created for easy reference when writing the model. In order to avoid errors caused by directly increasing parameters from 0 to simulated values ​​in the initial stage of model operation, in the runoff simulation, this embodiment sets 2010-2013 as the warm-up period, 2014-2018 as the rate period, and 2019-2023 as the verification period to complete the establishment and operation of the SWAT model of the Bosten Lake Basin.

[0104] By improving the hydrological and water temperature modules in the SWAT model (refer to the previous examples for details), a distributed hydrothermal coupling model for the Bosten Lake basin was developed. The model was calibrated using measured runoff data from the Dashankou hydrological station from 2014 to 2018 and validated using measured runoff data from the station from 2019 to 2023. The simulation results are shown in Table 1. The results demonstrate good simulation performance and are able to accurately describe the runoff process in the Bosten Lake basin.

[0105] Table 1 Runoff calibration and verification results

[0106]

[0107] The model was calibrated using measured water temperature data from the Dashankou hydrological station from 2014 to 2018 and validated using measured water temperature data from the Dashankou station from 2019 to 2023. The simulation results are shown in Table 2. The results show that the simulation effect is good and can basically accurately describe the water temperature process in the Bosten Lake basin.

[0108] Table 2 Water temperature calibration and verification results

[0109]

[0110] The specific process of determining and screening the predictive variables of Xinjiang naked lip fish is specifically referred to the above embodiment.

[0111] The point data of Xinjiang naked lip fish in the Bosteng Lake basin were collected, specifically including: analysis based on the results of two fish eDNA surveys in April 2023 and July 2023, with the unique fish Xinjiang naked lip fish as the target indigenous fish for analysis, and 25 distribution point records were identified. In order to avoid the model prediction deviation caused by multiple distribution sites in the same grid, the "spThin" package was called in the R software to perform sparse processing on the original distribution sites by retaining only one distribution point per square kilometer to reduce the impact of spatial autocorrelation on model prediction. Finally, independent distribution points were retained for subsequent species distribution model construction. A relative abundance distribution model of Xinjiang naked lip fish was constructed, and the relative abundance distribution model was used to simulate and predict the relative abundance of Xinjiang naked lip fish in the future. The process is specifically referred to the above embodiment, and the final result is as follows. Figure 2 shown.

[0112] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0113] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature, characterized in that: The specific steps are as follows: Constructing a watershed-distributed hydrothermal coupling model for the target area, and determining the hydrothermal environmental variables of the target area based on the watershed-distributed hydrothermal coupling model; determining and screening predictor variables related to target indigenous fish, wherein the predictor variables include the hydrothermal environment variables; The point data of target indigenous fish in the target area are collected, and a relative abundance distribution model of the target indigenous fish is constructed by combining the screened prediction variables; the relative abundance distribution model is used to simulate and predict the relative abundance of the target indigenous fish in the future.

2. The method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature according to claim 1, characterized in that: The construction of a distributed water-heat coupling model for the target area specifically includes: Construct a spatial database based on the digital elevation data, land use type data, soil type data and glacier vector data of the target area; Construct an attribute database based on soil attribute data, meteorological data, runoff and water temperature observation data of the target area; Establishing a SWAT model based on the spatial database and the attribute database; The hydrological module and the water temperature module in the SWAT model are improved to obtain the basin distributed water-heat coupling model.

3. The method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature according to claim 1, characterized in that: The improvement of the hydrological module is as follows: adding a glacier module to the hydrological module; The water balance equation of the hydrological module is as follows: Among them, SW t is the soil moisture content at the end of day t; SW0 is the initial soil moisture content; Pr i , Q surf,i 、E i , I i , Q lat,i and Q gr,i are the precipitation, surface runoff, evapotranspiration, soil infiltration, intersoil flow and base flow on day i, all in mm; The glacier melting equation of the glacier module is as follows: W′ g -W g =-(1-f)M+F-S; Among them, W g is the initial water equivalent, W g ' is the updated water equivalent, M is the glacier melt volume, f is the re-solidification ratio of glacier meltwater, f=0.2, F is the accumulated amount of glacier material, and S is the sublimation rate of the glacier; the glacier melt volume M is equivalent to precipitation and superimposed on the precipitation in the water balance equation, thereby adding the simulation of glacier runoff to the water cycle simulation and obtaining an improved hydrological module.

4. The method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature according to claim 1, characterized in that: The improvement of the water temperature module is: the temperature calculation formula of the original water temperature module is replaced by a water temperature calculation formula that takes into account the influence of different runoff components on water temperature. The formula is: Among them, T water is the water temperature, j represents the sub-basin number; T air is the average daily temperature; KK is the overall heat transfer coefficient, slr is the solar radiation; TT is the time it takes for water to pass through the sub-basin; η is an additional parameter for the water-air heat transfer process; ρ w is the density of water, is the specific heat capacity of water, H is the water depth; Twater intial is the temperature of the upstream sub-basin.

5. The method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature according to claim 4, characterized in that: The calculation formula for the upstream sub-basin temperature is: Among them, T waterupstrem is the water temperature of the river entering the sub-basin; Q outlet is the outlet flow of the sub-basin; sub_wyld is the total water production of the sub-basin; T w,local is the local water temperature of the sub-basin, and the calculation formula is: Among them, T w,local represents the local water temperature of the sub-basin; sub_snow is the snowmelt contribution of the sub-basin, sub_gla is the glacier melt contribution of the sub-basin, sub_gw is the groundwater contribution of the sub-basin, sub_surq is the surface runoff contribution of the sub-basin, and sub_latq is the lateral soil flow of the sub-basin; α, β, λ, and κ are the calibration coefficients of the relative contribution of each runoff component to the local water temperature; T sm represents the snowmelt temperature, T gw represents the groundwater temperature, T gla represents the glacier melt runoff temperature, T air,lag represents the lagged daily average temperature, and the calculation formula is: T air,lag,i =d×T air,lag(i-1) +(1-d)T air,i ; Where d is the hysteresis coefficient, T air is the average daily temperature, and i represents the number of simulation days.

6. The method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature according to claim 1, characterized in that: The determination and screening of predictor variables related to target indigenous fish species specifically include: Determine the bioclimatic variables, human activity variables, physical geographical variables and hydrothermal environment variables of the target area; Then, Spearman correlation coefficient analysis was performed on each variable, and environmental variables with correlation coefficient r ≥ |0.8| were eliminated; Variance inflation factors of the remaining variables were calculated, and predictor variables with variance inflation factors of a higher order of magnitude than other variables were eliminated in each round. The remaining variables were used to model the target indigenous fish species.

7. The method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature according to claim 1, characterized in that: The construction of the relative abundance distribution model of target indigenous fish species specifically includes: Collect point data of target indigenous fish in the target area and perform sparse processing; Based on the sparsely processed point data and the screened predictor variables, the maximum entropy model was used to predict the suitable habitat of the target indigenous fish. Based on the sparsely processed point data and the screened predictor variables, a generalized linear mixed model was used to simulate the abundance distribution of target indigenous fish.

8. The method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature according to claim 7, characterized in that: The regulation ratio and feature combination of the maximum entropy model were adjusted by the AIC information criterion, and the regulation ratio value and feature combination parameters with the smallest AIC value were selected to construct the final model.

9. The method for simulating the distribution of indigenous fish species coupled with hydrology and water temperature according to claim 7, characterized in that: The formula for the generalized linear mixed model is: y=Xτ+Zu+e; Among them, y is the vector of observed values ​​of fish relative abundance, τ is the vector of all fixed effects, X is the correlation matrix of fixed effects, u is the vector of all random effects, Z is the correlation matrix of random effects, and e is the random residual vector.

Images