Carbon nitrogen water coupling simulation method based on hydrological response unit

Abstract

The invention relates to a carbon nitrogen water coupling simulation method based on a hydrological response unit. The method comprises the following steps: collecting and processing data; an SWAT model is constructed; calibrating hydrological parameters; preparing an input file; the DNDC model is operated; evaluating a model effect; and simulating the carbon-nitrogen water flux. According to the method, the HRU serves as the minimum simulation unit of the DNDC model, and compared with grid simulation, calculation resources are saved, and the visualization effect is optimized. By constructing the research area SWAT model and calibrating the hydrological parameters, more accurate hydrological parameters are obtained, the calibrated hydrological parameters serve as input of the DNDC model, and the simulation precision of the DNDC model on hydrological variables is improved. In general, the DNDC model is applied to the watershed scale by taking the HRU as the minimum simulation unit, so that the simulation precision of the model on the hydrological process is improved, the calculation resources are saved, and the calculation speed is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to a carbon-nitrogen-water coupling simulation method based on hydrological response unit, relates to the cross field of ecological hydrology, hydrological model and biogeochemical process model, and is particularly suitable for comprehensive analysis of climate change, grazing and soil carbon-nitrogen-water cycle in grassland watershed ecosystem. BACKGROUND

[0002] Under the background of global climate change, agricultural management needs to balance between environmental protection and economic benefits, especially in grazing systems, the dual consideration of carbon sequestration and carbon emission is also increasingly important. Agricultural ecosystem models, such as DNDC (Denitrification-Decomposition), APSIM (Agricultural Production Systems sIMulator), DayCent (Daily Century Model) and Roth C (Rothamsted Carbon Model), are mainly used to simulate the impact of agricultural management on soil carbon and nitrogen cycle. Most of these models are based on point-scale driving, and it is difficult to simulate regional-scale hydrological processes such as surface runoff and nutrient loss. Due to the limitation of code structure, many traditional models cannot effectively couple hydrological processes, and it is difficult to simulate stream, nutrient migration and leakage processes. Therefore, in large-scale research, how to integrate agricultural ecosystem model and watershed hydrological model is a problem to be solved. SUMMARY

[0003] In order to overcome the problems of the prior art, the present application proposes a carbon-nitrogen-water coupling simulation method based on hydrological response unit. The method couples hydrological model and biogeochemical process model, aiming to improve the simulation accuracy of watershed carbon-nitrogen-water cycle, mainly in the improvement of DNDC model's simulation ability for watershed hydrological process, and can realize easy acquisition of hydrological parameter value, saving of calculation amount and calculation resource.

[0004] The purpose of the present application is achieved by a carbon-nitrogen-water coupling simulation method based on hydrological response unit, the steps of the method are as follows:

[0005] Step 1, data collection and processing: obtain meteorological hydrological data, soil, land use type, elevation and management measures of the research area, respectively, pre-process the collected data, interpolate the missing values of meteorological hydrological data, and extract and process remote sensing data to the same resolution;

[0006] Step 2, building a SWAT model: build a SWAT model of the study area, reclassify the land use and soil data according to the actual situation of the study area; determine the number of hydrologic response units according to the classification of land use and soil; wherein, in the process of defining HRU, no merging operation is performed on HRU, so as to ensure that each HRU is of the same land use type, soil type and slope;

[0007] Step 3, hydrological parameter calibration: calibrate the model parameters according to the existing long series of hydrological variables, wherein the conventional hydrological variables are runoff, soil water content or evaporation, and the Nash efficiency coefficient NSE and the determination coefficient R 2 as the objective function to evaluate the simulation accuracy of the model;

[0008] Step 4, preparing input files: obtaining hydrologic response unit information from the built SWAT model, taking the HRU generated in step 2 as the minimum simulation unit of the DNDC model to prepare the DNDC model input file, inputting the localized model parameters, and importing the hydrological parameter values calibrated in step 3, forming multiple DNDC model input files;

[0009] Step 5, running the DNDC model: batch running the multiple DNDC model input files generated in step 4 in the DNDC model to simulate the carbon-nitrogen-water cycle of the study area, simultaneously performing sensitivity analysis on the model parameters, and parameterizing the model according to the measured values;

[0010] Step 6, model effect evaluation: extracting the DNDC carbon-nitrogen-water cycle flux simulation results, evaluating the model effect and performance according to multiple index values, and selecting the determination coefficient R 2 , root mean square error RMSE, and normalized root mean square error nRMSE to evaluate the simulation accuracy and performance of the DNDC model;

[0011] Step 7, carbon-nitrogen-water flux simulation: simulating carbon-nitrogen-water fluxes using the HRU-based DNDC model, including soil water, net primary productivity NPP, total nitrogen TN, soil organic carbon SOC and other fluxes.

[0012] The present application has the advantages and beneficial effects that: the present application uses a method of constructing a watershed SWAT model, obtaining hydrological parameter values on the HRU scale, and taking HRU as the input of the DNDC model, coupling the SWAT model and the DNDC model, improving the simulation accuracy of the DNDC model on the watershed hydrological process, and thus more accurately simulating the carbon-nitrogen-water cycle flux. The present application takes HRU as the minimum simulation unit of the DNDC model, which saves computing resources and optimizes the visualization effect compared to grid simulation. By constructing the SWAT model of the research area and calibrating the hydrological parameters, more accurate hydrological parameters are obtained, and the calibrated hydrological parameters are taken as the input of the DNDC, which improves the simulation accuracy of the DNDC model on the hydrological variables. In summary, the present application applies the DNDC model to the watershed scale with HRU as the minimum simulation unit, improves the simulation accuracy of the model on the hydrological process, saves computing resources, and improves the calculation speed. BRIEF DESCRIPTION OF DRAWINGS

[0013] The present application will be further described below in combination with the drawings and examples.

[0014] Figure 1 is a flow chart of the method described in the embodiments of the present application;

[0015] Figure 2 is a verification diagram of the carbon-nitrogen-water coupling simulation of the research area Xilinha River Basin of the application example described in the embodiments of the present application;

[0016] Figure 3 is a schematic diagram of the change and distribution of net primary productivity NPP of the carbon-nitrogen-water coupling simulation of the research area Xilinha River Basin of the application example described in the embodiments of the present application. DETAILED DESCRIPTION

[0017] Embodiment:

[0018] DNDC is a biogeochemical model that simulates carbon and nitrogen dynamics at daily time scale, in which a soil thermal-hydraulic module is used to evaluate the daily variation of soil temperature, oxygen and water, and a plant growth sub-model is used to consider the water and nutrient uptake by vegetation. In the original version of DNDC, a one-dimensional soil water flow model is used to calculate the water dynamics of soil profile, and in the improved version of DNDC model, a hydrological module is introduced to simulate the water flow process and soil nitrogen leaching simultaneously by using SCS curve and MUSLE equation, and the required hydrological parameters are generally obtained from experiments or estimation. DNDC model can be driven in two ways, i.e. point and regional simulation, and the regional simulation can be regarded as uniformly distributed point simulation. The SWAT (Soil and Water Assessment Tool) model is widely used in the fields of hydrology, soil and agricultural management, and the model divides the study area into multiple hydrologic response units (HRU, Hydrologic Response Unit) according to different soil, land use and slope, which are used as the minimum simulation unit of the model.

[0019] The limitation of applying DNDC model to simulate carbon, nitrogen and water cycle at watershed scale is that it is difficult to obtain accurate hydrological parameter values through experiments or estimation. Considering the driving mode of the model is point or regional simulation, in order to improve the simulation accuracy, it is necessary to divide more grids, i.e. a large number of points for simulation, and the spatial heterogeneity of soil, land use, slope and other factors brings great difficulty and uncertainty to the acquisition of hydrological parameter values. The limitation of using SWAT model to simulate carbon, nitrogen and water cycle is that the description of soil carbon and nitrogen cycle in SWAT model itself is relatively simple, and a key factor of soil carbon and nitrogen cycle is the activity and effect of microorganisms, but the SWAT model does not simulate the influence of microorganisms on soil organic matter decomposition, nitrogen transformation and other processes in detail.

[0020] In order to more accurately and efficiently simulate the carbon, nitrogen and water cycle of the watershed, an embodiment of the present application proposes a carbon and nitrogen water coupling simulation method based on hydrologic response unit, which acquires hydrological parameter values at HRU scale by constructing a SWAT model of the watershed, and uses HRU as the input of DNDC model, and couples the SWAT model and the DNDC model to improve the simulation accuracy of the DNDC model on the hydrological process of the watershed, so as to more accurately predict the carbon, nitrogen and water cycle flux. The present embodiment will be described in detail in combination with application examples, and the method comprises the following steps:

[0021] Step 1, data collection and processing: obtain the meteorological and hydrological data, soil data, land use type, elevation data and management measure data of the study area, respectively pre-process the collected data, interpolate the missing values of meteorological and hydrological data, and extract the remote sensing data to the same resolution.

[0022] Application Example:

[0023] This application example takes the Xilin River Basin in Inner Mongolia as the research area, collects daily scale meteorological and hydrological data of the Xilin River Basin from 1960 to 2019, including runoff data and temperature data, i.e. precipitation, temperature, wind speed, radiation, humidity, etc., interpolates the missing values, and formats them into standard meteorological data input formats of the SWAT model and the DNDC model; collects soil type data, land use type data and elevation data, and extracts them to the same resolution; collects grazing and mowing related parameters of the Xilin River Basin from 1980 to 2019, including time and intensity, as well as related attribute parameters of livestock.

[0024] Step 2, build the SWAT model: build the SWAT model of the research area, reclassify the land use and soil data according to the actual situation of the research area. Determine the number of hydrological response units according to the classification of land use and soil. In the HRU definition process, do not merge the HRUs to ensure that each HRU has the same land use type, soil type and slope.

[0025] Based on the collected basic data, the SWAT model of the Xilin River Basin is built, wherein the reclassification of land use and soil data according to the actual situation of the Xilin River Basin is as follows: the land use data is divided into nine categories, namely water area, wetland, urban land, wasteland, agricultural land, shrub land, low-covered grassland, medium-covered grassland and high-covered grassland. The soil data is divided into five categories: lime black calcic soil, leaching chestnut calcic soil, lime red sandy soil, gley black soil and calcium chestnut soil. According to the above classification, a total of 466 hydrological response units are generated, of which 262 hydrological response units are grassland. In the definition step of the hydrological response unit, the HRUs are not merged, which is to ensure that the final hydrological response unit has the same land use type, soil type and slope.

[0026] Step 3, hydrological parameter calibration: calibrate the model parameters according to the existing long series of hydrological variables, wherein the calibration of the hydrological parameter value can be realized by automatic calibration through the SWATCUP software or other software, or by manual calibration. The conventional hydrological variables are runoff, soil water content or evaporation, and the Nash efficiency coefficient NSE and the determination coefficient R 2 are selected as the objective function to evaluate the simulation accuracy of the model.

[0027] The calculation formula of the Nash efficiency coefficient NSE is as follows:

[0028]

[0029] The calculation formula of the determination coefficient R 2 is as follows:

[0030]

[0031] where y i is the observed value; is the simulated value; is the mean of the observed values; is the mean of the simulated values; n is the sample size, (1, 2, …, i, …, n).

[0032] The daily runoff data of the total export Xilinhot hydrological station of Xilin River Basin is used for rating the SWAT model. The runoff of the total export of the basin is automatically rated by the SWATCUP software for model parameters. In the rating process, the step of repeatedly changing the parameter interval of the model and running the program is needed until the evaluation index meets the requirements.

[0033] After rating, when the Nash efficiency coefficient NSE > 0.5 and the coefficient of determination R 2 > 0.6 in the study area Xilin River Basin, it is considered that the SWAT model has good simulation effect on Xilin River Basin at this time, and the best value of each model parameter is recorded.

[0034] Step 4, preparation of input file: obtain the hydrological response unit information from the constructed SWAT model, take the HRU generated in step 2 as the minimum simulation unit of DNDC model to prepare the DNDC model input file, input the localized model parameters at the same time, and import the hydrological parameter values rated in step 3, forming multiple DNDC model input files.

[0035] In the study area Xilin River Basin, the focus is on simulating the fluxes of carbon and nitrogen water cycle in the grassland area, so only the 262 hydrological response units of grassland land use type generated in step 2 are extracted for the preparation of DNDC input file. According to the point simulation input file template of DNDC model, the soil parameters, meteorological parameters, input, vegetation parameters and management measures parameters are set in turn. The hydrological module is opened, and the hydrological parameter values of each hydrological response unit are imported, which are: CN2, OV_N, CH_N2, HRU_SLP, SLSUBBSN corresponding to SCS_curve_number, land_surface_roughness, channel_surface_roughness, channel_slope, channel_length respectively.

[0036] Step 5, running DNDC model: batch running multiple DNDC model input files generated in step 4 in DNDC model to simulate carbon and nitrogen water cycle in the study area, at the same time, sensitivity analysis of model parameters is carried out, and the model is parameterized according to the measured value.

[0037] In the study area of Xilin River Basin, the simulation time of DNDC model is from 1960 to 2019. Batch run 262 DNDC input files prepared in step 4 to simulate carbon and nitrogen cycle in the study area of Xilin River Basin. In the simulation process, the parameters of the model are analyzed and parameterized. Because DNDC model does not have a special tool for parameter adjustment, therefore, the parameterization of the model needs to rely on the sensitivity analysis of the model to adjust the model parameters. The specific way is to adjust the model parameters according to the difference between the simulation value and the measured value.

[0038] Step 6, model effect evaluation: extract the simulation results of DNDC carbon and nitrogen cycle flux, evaluate the model effect and performance according to multi-index value, the index selection includes determination coefficient R 2 , root mean square error RMSE, normalized root mean square error nRMSE, through the three indexes to evaluate the simulation accuracy and performance of DNDC model.

[0039] The calculation formula of determination coefficient R 2 is as follows:

[0040]

[0041] The calculation formula of root mean square error RMSE is as follows:

[0042]

[0043] The calculation formula of normalized root mean square error nRMSE is as follows:

[0044]

[0045] Where, y i is the observed value; is the simulation value; is the average value of the observed value; is the average value of the simulation value; n is the sample number, (1, 2, …, i, …, n).

[0046] In the study area of Xilin River Basin, the soil water content SW (Soil Water) from 2000 to 2019 and the net primary productivity NPP (Net Primary Productivity) from 2001 to 2019 were extracted from the DNDC output for model accuracy verification (the selection period is limited by the length of the observation data). The soil water for verification is the SMCI1.0 dataset, and the net primary productivity data for verification is the MOD17A3HGF Version 6.1 dataset (https: / / lpdaac.usgs.gov / products / mod17a3hgfv061 / ). The net primary productivity is verified in space and time. Since the conversion between soil water content and soil moisture involves soil texture, only the soil water is verified in time. The three indicators comprehensively represent that the simulation effect of DNDC model on net primary productivity and soil water is good. Spatially, for the simulation verification of NPP, the determination coefficient R 2 reaches 0.89, the root mean square error RMSE is 310, and the normalized root mean square error nRMSE is 0.17. Temporally, the R 2 of NPP is between 0.4 and 0.8, and the R 2 of soil water is between 0.5 and 0.7. See Figure 2 . Figure 2

[0047] a: The selection of spatial verification points ensures uniform distribution and avoids water areas.

[0048] b: Spatial verification results of net primary productivity NPP in 2019.

[0049] c: Verification results of net primary productivity NPP on hydrological response unit HRU scale from 2001 to 2019.

[0050] d: Verification results of soil water SW on hydrological response unit HRU scale from 2000 to 2019.

[0051] Step 7, simulation of carbon, nitrogen and water fluxes: DNDC model based on HRU is used to simulate carbon, nitrogen and water fluxes, including but not limited to soil water SW (Soil Water), net primary productivity NPP (Net Primary Productivity), total nitrogen TN (Total Nitrogen), soil organic carbon SOC (Soil Organic Carbon), etc.

[0052] In the study area of Xilin River Basin, the trend and characteristics of net primary productivity from 1985 to 2019 were simulated and analyzed by DNDC model (1960-1980 as model preheating period), see​Figure 3 The net primary productivity in the northeast and southeast of Xilin River Basin is higher, and the central part is lower, which is consistent with the actual situation of Xilin River Basin. The northeast and southeast of Xilin River Basin are typical grassland regions, and the dominant species are Leymus chinensis and Stipa grandis. The central part is distributed with desert steppe, and the dominant species are Caragana microphylla and the like. In terms of time, the net primary productivity of Xilin River Basin reaches the highest level around 1990, and is greatly reduced after 2000, and gradually increases after 2010.

[0053] Finally, it should be noted that the above is only used to illustrate the technical solutions of the present application, not to limit. Although the present application is described in detail with reference to the preferred arrangement, those skilled in the art should understand that the technical solutions of the present application (such as the collection method of data, the processing method of different software, the use of various formulas, the order of steps, etc.) can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A carbon-nitrogen-water coupling simulation method based on a hydrological response unit, characterized in that: The steps of the method are as follows: Step 1, data collection and processing: obtain meteorological and hydrological data, soil data, land use type, elevation data, and management measures data of the study area, pre-process the collected data, interpolate missing values ​​of meteorological and hydrological data, and extract remote sensing data to the same resolution; Step 2: Construct a SWAT model: Construct a SWAT model for the study area and reclassify the land use and soil data according to the actual conditions of the study area. Determine the number of generated hydrological response units based on the land use and soil classifications. HRUs are not merged during the HRU definition process to ensure that each HRU has the same land use type, soil type, and slope. Step 3: Calibrate the hydrological parameters: calibrate the model parameters based on the existing long series of hydrological variables, where the conventional hydrological variables are runoff, soil water content or evaporation, and select the Nash efficiency coefficient NSE and the coefficient of determination R 2 As an objective function to evaluate the model simulation accuracy; Step 4, prepare input files: obtain hydrological response unit information from the constructed SWAT model, use the HRU generated in step 2 as the minimum simulation unit of the DNDC model to prepare the DNDC model input file, input the localized model parameters, and import the hydrological parameter values ​​calibrated in step 3 to form multiple DNDC model input files; Step 5, run the DNDC model: batch run multiple DNDC model input files generated in step 4 in the DNDC model to simulate the carbon, nitrogen and water cycles in the study area, perform sensitivity analysis on the model parameters, and parameterize the model according to the measured values; Step 6, model effect evaluation: extract the DNDC carbon, nitrogen and water cycle flux simulation results, and evaluate the model effect and performance based on multiple index values. The index selection includes the determination coefficient R 2 , root mean square error RMSE , normalized root mean square error nRMSE ,These three indicators are used to evaluate the simulation accuracy and performance of the DNDC model; Step 7, carbon, nitrogen and water flux simulation: Use the HRU-based DNDC model to simulate carbon, nitrogen and water fluxes, including fluxes of soil water, net primary productivity (NPP), total nitrogen (TN), soil organic carbon (SOC), etc.

Images