A method and system for predicting water and fertilizer demand for increasing crop yield

Abstract

The present invention provides a method and system for predicting water and fertilizer demand for improving crop yields, and relates to a data processing system, including: S1, obtaining meteorological data; S2, calculating soil water content data; S3, recording farming data; S3, calculating soil fertility data; S5, setting a root depth calculation module and a pollination rate calculation module in a Wofost model, and optimizing the Wofost model according to the calculation results of the root depth calculation module and the pollination rate calculation module; S6, based on the Wofost model, inputting meteorological data, soil water content data, soil fertility data, and farming data to perform yield estimation calculations; S7, calculating daily irrigation demand according to meteorological data, soil water content data, farming data, and crop transpiration data; S8, calculating daily fertilization demand according to soil fertility data, farming data, and nutrient demand data. The system can help farmers prepare before abnormal water and fertilizer conditions occur in farmland, realize accurate management of water and fertilizer, reduce water and fertilizer waste, and improve the yield and quality of agricultural products.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing systems, and in particular to a method and system for predicting water and fertilizer demands for increasing crop yields. Background Art

[0002] In the crop yield improvement project, water and fertilizer management is the top priority. In terms of agricultural management, the yield improvement project should first summarize the planting management experience of previous years, examine the omissions in planting management, and find active response strategies; second, it should timely discover the water and fertilizer needs of crops, pests and diseases, and extreme weather in the current planting management, and prepare response measures. In summarizing the planting management experience of previous years, the effectiveness of water and fertilizer management in previous years cannot be accurately judged by traditional planting methods, but with the gradual development of crop models, it has been possible to effectively simulate the crop growth process and effectively guide planting operations. The crop model (crop growth model) is based on the crop physiological process mechanism, and takes the factors affecting crop growth such as climate, soil, crop varieties and management measures as a numerical simulation system of an overall system. It can dynamically simulate the biological parameters of crop growth and development at a single point scale and crop yield, and quantitatively study the effects of environmental factors and field management measures on crop growth and development.

[0003] At present, the internationally used crop models are mainly divided into remote sensing models and computational models. Among them, remote sensing models have been put into commercial application, while computational models are mainly used in the scientific research stage. At present, the commonly used crop computational models in the world are DSSAT in the United States, Apsim in Australia, and Wofost in the Netherlands. The reasons why computational models have not been widely used are: 1. Various parameters in the model are obtained through measurement, and it is difficult for non-research institutions to control such operations; 2. The model content and parameters are too academic and too professional to be practical for the public; 3. These models are used by scientific research institutes, and some parameter settings cannot fit actual production; 4. The computational model pays more attention to the results rather than the process in the application, so the current models are all post-calculation models; 5. The models are all fixed models, without self-improvement mechanisms, and rely on subsequent in-depth development and improvement; 6. The models can only input data manually, are not suitable for web applications, and have no external interfaces. For the crop yield improvement project, the above models cannot be directly and effectively applied by farmers, and do not have high accuracy and practicality for the water and fertilizer demand prediction task that farmers (growers) are most concerned about.

[0004] Therefore, how to design a prediction system with a low operating threshold that can accurately predict the water and fertilizer needs of crops has become a technical problem that needs to be solved. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art or related technology, and discloses a method and system for predicting water and fertilizer demand for improving crop yields. After receiving planting and management parameters input by farmers, variable calculations are performed to help farmers review agricultural production management and facilitate the inspection of planting management omissions.

[0006] The first aspect of the present invention discloses a method for predicting water and fertilizer demand for improving crop yields, comprising: S1, obtaining meteorological data, the meteorological data including: temperature, wind speed, evaporation, precipitation and solar radiation, and inputting the obtained meteorological data into a Wofost model; S2, calculating soil water content data, the soil water content data including: saturated water content, field water holding capacity, and wilting point, and the meteorological data is used to calculate the soil water content data; S3, recording agricultural data, the agricultural data including: farmland area, planting variety, planting date, planting amount, irrigation date, irrigation amount, fertilization date, fertilization amount, and fertilization type; S4, calculating soil fertility data, the soil fertility data including: nitrogen content, phosphorus content, potassium content, mineralization data, and fertilizer loss data, and the agricultural data is used to calculate the soil fertility data; S5, setting a root depth calculation module and a pollination rate calculation module in the Wofost model, and optimizing the Wofost model according to the calculation results of the root depth calculation module and the pollination rate calculation module, wherein the root depth calculation module is used to calculate the soil fertility data. The actual root depth data of crops is calculated, and the soil water content data and soil fertility data corresponding to the soil layer touched by the roots are selected according to the actual root depth data to participate in the yield estimation calculation process of the Wofost model, so as to obtain the water and fertilizer demand calculation results and the yield estimation calculation results. The pollination rate calculation module is used to calculate the pollination rate of crops, and to correct the water and fertilizer demand calculation results output by the Wofost model according to the pollination rate, and to correct the yield estimation calculation results according to the pollination rate; S6, based on the optimized Wofost model, meteorological data, soil water content data, soil fertility data, and farming data are input to perform yield estimation calculation, and the yield estimation calculation result with the highest yield and the corresponding water demand data, nutrient demand data, and crop transpiration data are selected, wherein the nutrient demand data is the nutrient demand calculated by the Wofost model in the unlimited water mode; S7, daily irrigation demand is calculated according to meteorological data, soil water content data, farming data, and crop transpiration data; S8, daily fertilization demand is calculated according to soil fertility data, farming data, and nutrient demand data.

[0007] Furthermore, it also includes: by repeatedly modifying the irrigation date, irrigation amount, fertilization date, fertilization amount or fertilization type for calculation, the Wofost model outputs multiple sets of crop growth status data and yield data to determine the growth status and yield of crops under different conditions, thereby obtaining the best planting date and the best harvest date.

[0008] Furthermore, S2 specifically includes: according to the soil survey data, the soil layer is divided into 8 depths: 4.5 cm, 9.1 cm, 16.6 cm, 28.9 cm, 49.3 cm, 82.9 cm, 138.3 cm and 229.6 cm; the sand, loam, clay, bulk density and organic matter content data in the soil are used as the basic data; the soil texture data, texture content data, time series data and precipitation in the meteorological data are combined and substituted into the VanGenuchten water characteristic curve equation; the soil saturated moisture content, field water holding capacity and wilting point in the soil are calculated in combination with the VanGenuchten water characteristic curve parameters; and the soil water content data is calculated for each depth of the soil layer.

[0009] Furthermore, S4 specifically includes: according to the soil census data, the soil layer is divided into 8 depths: 4.5 cm, 9.1 cm, 16.6 cm, 28.9 cm, 49.3 cm, 82.9 cm, 138.3 cm and 229.6 cm; according to the effective nitrogen, phosphorus and potassium, total nitrogen, total phosphorus, total potassium and organic matter content in the soil and the fertilization date, amount and type of fertilization in the agricultural data, the nitrogen content, phosphorus content, potassium content, mineralization data and fertilizer loss data are calculated; and the soil fertility data are calculated for each depth of the soil layer.

[0010] Furthermore, the calculation process of the root depth calculation module specifically includes:

[0011] Establish effective root depth function and root density function;

[0012] According to the accumulated temperature and the root nutrient distribution function during the growth stage, the daily root dry matter accumulation and daily effective root depth were calculated;

[0013] According to the daily effective root depth calculation results, the actual depth reached by the roots is determined to obtain the actual root depth data. Based on the actual root depth data, the soil water content data and soil fertility data corresponding to the soil layer touched by the roots are selected as the input data of the Wofost model.

[0014] Furthermore, the calculation process of the pollination rate calculation module specifically includes:

[0015] Crop pollination is classified according to self-pollination and alienation pollination, and different pollination rates are preset for different temperatures during pollination, different pollination rates are preset for different humidity during pollination, different pollination rates are preset for different wind speeds during pollination, and different pollination rates are preset for different precipitation intensities during pollination;

[0016] According to the temperature, humidity, wind speed, and precipitation intensity in the meteorological data, the four pollination rates at the corresponding time are obtained, the lowest value of the four pollination rates is taken, and the pollination rates per hour are accumulated to obtain the final hourly probability of pollination during the flowering period;

[0017] Pollination rate = the final hour probability of pollination during flowering period / (number of days of crop flowering period*24 hours);

[0018] If the pollen amount of crops is set to be 3 times the required amount, then 1 / 3 of the pollen is enough for 100% pollination of the crops;

[0019] The adjusted pollination rates are:

[0020] Pollination rate = the final hour probability of pollination during flowering period / (number of days of crop flowering period * 24 hours) / (1 / 3);

[0021] The pollination rate is multiplied by the water and fertilizer demand calculation results and the estimated yield calculation results output by the Wofost model to obtain the revised water and fertilizer demand calculation results and the estimated yield calculation results.

[0022] Furthermore, the training process of the Wofost model includes:

[0023] The planting management data and yield data from multiple locations and multiple growing seasons are collected in a mixed manner, and satellite remote sensing data within the growing season is obtained at the same time. The remote sensing data is coupled with meteorological data, soil water content data, farming data, and soil fertility data, and the Wofost model is obtained through deep learning.

[0024] Furthermore, it also includes: storing and / or reading the estimated production calculation process data and the estimated production calculation results of the Wofost model on a daily basis to support breakpoint insertion of remote sensing data and API data and perform breakpoint calculation.

[0025] The second aspect of the present invention also discloses a water and fertilizer demand prediction system for improving crop yields, comprising: a memory for storing program instructions; a processor for calling the program instructions stored in the memory to implement a water and fertilizer demand prediction method for improving crop yields as provided in any of the above technical solutions.

[0026] The beneficial effects of the present invention include at least: the present application improves the calculation process of the Wofost model, and improves the water and fertilizer calculation accuracy of the Wofost model by adding a root depth calculation module and a pollination rate calculation module; the technical solution disclosed in the present application can perform variable calculations after farmers input planting and management parameters, helping farmers to review agricultural production management, conveniently review planting management omissions, and find active response strategies. As an irrigation and fertilization forecast model during the growth of crops, farmers can prepare agricultural activities for the next two weeks very conveniently. The data output every day can make farmland fertilization and irrigation more accurate, which can avoid waste of water and fertilizer, reduce input in agricultural production, and save resources and protect the environment. If farmers can use data for drip irrigation or sprinkler irrigation control, precise control of water and fertilizer can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A flow chart of a method for predicting water and fertilizer requirements for improving crop yields according to an embodiment of the present invention is shown.

[0028] Figure 2 A schematic block diagram of a water and fertilizer demand prediction system for improving crop yield according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0029] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0031] like Figure 1 As shown, according to one embodiment of the present invention, a method for predicting water and fertilizer demand for improving crop yield is disclosed, comprising:

[0032] S1, obtain meteorological data: download national daily meteorological data from NASA and the European Meteorological Center, including temperature, wind speed, evaporation, precipitation, and solar radiation, etc., to support soil water content-related calculations and soil fertility calculations, and as input meteorological data for the Wofost model to participate in the calculation; specific data include: maximum temperature, minimum temperature, wind speed, total precipitation, surface runoff, air pressure, surface net shortwave radiation, groundwater runoff, surface water retention, canopy water dew, canopy water layer evaporation, water surface evaporation, bare soil evaporation, and water content in different soil layers. Among them, the maximum temperature, minimum temperature, wind speed, total precipitation, air pressure, and surface net shortwave radiation data are used as wofost input meteorological data; other data are used in the final calculation of water and fertilizer.

[0033] S2, calculate soil water content data: soil water content data include: saturated water content, field water holding capacity, wilting point, and meteorological data are used to calculate soil water content data;

[0034] In this step, soil refers to the soil at the location of the farmland circled by the user during use; saturated moisture content refers to the moisture content when all soil pores are filled with water under natural conditions; field water holding capacity: refers to the relatively stable soil moisture content that can be maintained in the soil profile after sufficient irrigation or precipitation on land with deep groundwater and good drainage, allowing water to fully infiltrate and preventing its evaporation after a certain period of time; wilting point refers to the lowest soil moisture content in the soil that cannot be absorbed and utilized by plants. When the moisture content is lower than this, the plants will wilt and die.

[0035] S3, record agricultural data: agricultural data include: farmland area, planting varieties, sowing date, sowing amount, irrigation date, irrigation amount, fertilization date, fertilization amount, and fertilizer type;

[0036] In this step, the farmland area refers to the farmland area circled by the user when using it; planting varieties: the varieties that the user needs to calculate when using the system; planting date: the planting date corresponding to the growing season that the user needs to calculate when using the system; planting amount: the planting amount corresponding to the growing season that the user needs to calculate when using the system; irrigation date: the date of each irrigation in the growing season that the user needs to calculate when using the system; irrigation amount: the water layer thickness of each irrigation in the growing season that the user needs to calculate when using the system; fertilization date: the date of each fertilization in the growing season that the user needs to calculate when using the system; fertilization amount: the fertilization amount in the growing season that the user needs to calculate when using the system.

[0037] S4, calculate soil fertility data: soil fertility data includes: nitrogen content, phosphorus content, potassium content, mineralization data, fertilizer loss data, and agricultural data are used to calculate soil fertility data;

[0038] In this step, soil refers to the soil at the farmland site circled by the user during use; mineralization refers to the process of converting organic compounds in the soil into inorganic compounds under the action of soil microorganisms.

[0039] S5, optimizing the Wofost model: setting a root depth calculation module and a pollination rate calculation module in the Wofost model, and optimizing the Wofost model according to the calculation results of the root depth calculation module and the pollination rate calculation module, wherein the root depth calculation module is used to calculate the actual root depth data of the crops, and select the soil water content data and soil fertility data corresponding to the soil layer touched by the roots according to the actual root depth data to participate in the yield estimation calculation process of the Wofost model, so as to obtain the water and fertilizer demand calculation results and the yield estimation calculation results, and the pollination rate calculation module is used to calculate the pollination rate of the crops, and correct the water and fertilizer demand calculation results output by the Wofost model according to the pollination rate, and correct the yield estimation calculation results according to the pollination rate;

[0040] S6, model calculation: Based on the optimized Wofost model, input meteorological data, soil water content data, soil fertility data, and farming data to estimate yield. According to the data of the eight soil layers, the yield calculation results of each soil layer are estimated, and the yield calculation result with the highest yield and the corresponding water demand data, nutrient demand data, and crop transpiration data are selected. Among them, the nutrient demand data is the nutrient demand calculated by the Wofost model in the unlimited water mode;

[0041] In this step, the Wofost model (crop model) is an open source crop model developed by Wageningen University in the Netherlands; root depth calculation: crop root growth rate calculation, which is affected by weather, water and fertilizer, and also affects water and fertilizer absorption. Only when the root growth can reach the water and fertilizer can the water and fertilizer be absorbed normally; pollination rate calculation: crop pollination success percentage. By default, the main nutrients of the crop growth after flowering are supplied to the fruit part, and the maximum assimilation rate is restricted by the pollination rate percentage. Training: Based on the modified wofost model according to the root depth calculation and pollination rate calculation, deep learning is carried out with soil, previous weather, and previous agricultural management data as channels. Nutritional demand data: nitrogen, phosphorus and potassium requirements in a certain period of time, generally 7-15 days of nitrogen, phosphorus and potassium elements; crop transpiration data: the transpiration of water during crop growth is an indicator of crop growth water requirements.

[0042] S7, calculate daily irrigation demand: calculate daily irrigation demand based on farming data, soil water content data, crop transpiration data and meteorological data;

[0043] In this step, the agricultural data specifically include the calculation of irrigation date and irrigation amount data during the growth period, which are also used as the input data of agricultural data of the wofost model; soil water content data: soil saturated water content, field water holding capacity, and wilting point in S2, which are also used as soil input data in the wofost model; crop transpiration data: the transpiration amount of water during crop growth, which is obtained after calculation by the wofost model.

[0044] S7, calculate daily fertilizer demand: calculate daily fertilizer demand based on farming data, nutrient demand data and soil fertility data.

[0045] In this step, the agricultural data specifically include the calculation of the fertilization time and amount during the growth period, which are also used as the input data of the agricultural data of the wofost model; soil fertility data: soil nitrogen content, phosphorus content, potassium content, mineralization data, fertilizer loss data in S4, which are also used as soil input data of the wofost model; nutritional requirement data: nitrogen, phosphorus and potassium requirements during crop growth, which are obtained after calculation by the wofost model.

[0046] According to the above embodiment, the calculation input data of the wofost model include: S1 meteorological data, S2 soil water content data, S3 agricultural data, S4 soil fertility data. The calculation output data of the wofost model include: leaf area index LAI, dry matter accumulation, crop transpiration data, crop nitrogen, phosphorus and potassium demand data, and model-calculated soil moisture data. Among them, the leaf area index LAI and dry matter accumulation data are used for deep learning, the crop transpiration data are used for the final irrigation calculation, the crop nitrogen, phosphorus and potassium demand data are used for fertilization calculation, and the model-calculated soil moisture data will be averaged with the remote sensing data and applied to the final irrigation calculation.

[0047] According to the above embodiment, preferably, the calculation process of the root depth calculation module includes: establishing an effective root depth function and a root density function;

[0048] According to the accumulated temperature and the root nutrient distribution function during the growth stage, the daily root dry matter accumulation and daily effective root depth were calculated;

[0049] According to the daily effective root depth calculation results, the actual depth reached by the roots is determined to obtain the actual root depth data. Based on the actual root depth data, the soil water content data and soil fertility data corresponding to the soil layer touched by the roots are selected as the input data of the Wofost model.

[0050] The effective root depth and root density percentage functions are set, and the accumulated temperature and dry matter accumulation are used to gradiently calculate the root growth and development of crops, which are used to calculate the root depth and water and fertilizer reach of crops, and serve as the basic data for calculating water and fertilizer requirements.

[0051] In this embodiment, the root density function refers to the proportion of roots that can maintain the water and fertilizer supply of the plant under natural conditions and the highest assimilation rate of the crop. For example, at least 85% of the roots need to work to ensure normal growth of the crop under sunny weather conditions; assimilation: the carbon dioxide assimilation of the crop, that is, photosynthesis; effective depth: the maximum depth of the root system corresponding to the effective root density function, for example, in the above example, 85% of the roots grow in the soil layer above 20 cm; accumulated temperature: the accumulation of temperature required for crop growth, and the calculation method of accumulated temperature is different for different crop varieties; dry matter accumulation: the process of crop growth The organic matter generation process in the data of agricultural management; Taking the amount of seeding in the agricultural management data as the initial value, multiplying the daily assimilation rate in different periods by various impairment functions, and then accumulating them to obtain the dry matter accumulation data; Assimilation rate: that is, the assimilation speed. The basic data comes from experiments or literature, and is optimized and improved through deep learning; Impairment function: including temperature impairment function, water stress impairment function, fertilizer stress impairment function, light stress impairment function, etc. This type of data is the internal data of the wofost model, which comes from experiments or literature and is optimized and improved through deep learning; Water and fertilizer reach: the absorbable water and fertilizer that the roots can contact. Calculation method: Set the daily maximum root growth at the maximum assimilation rate of the crop, as well as the dry matter accumulation data. Based on the dry matter accumulation of the previous day, calculate the dry matter mass of dry matter allocated to the root, divide this dry matter mass by the daily maximum root growth, and then multiply it by the daily maximum root growth at the maximum assimilation rate of the crop to obtain the daily root depth growth data. Dry matter distribution: the proportion of dry matter distributed to roots, stems, leaves, and fruits at different times. The internal data of the wofost model comes from experiments or literature and is optimized and improved through deep learning.

[0052] According to the above embodiment, preferably, the calculation process of the pollination rate calculation module includes: classifying crop pollination according to self-pollination and alienated pollination, using meteorological factors such as temperature, humidity, wind speed, rainfall, etc. as pollination driving input items, calculating the effective pollination rate of the crop during flowering period, presetting different pollination rates for different temperatures during pollination, presetting different pollination rates for different humidity during pollination, presetting different pollination rates for wind speeds during pollination, and presetting different pollination rates for precipitation intensities during pollination; so that the crop model uses the effective pollination rate as a factor affecting water and fertilizer absorption and crop yield.

[0053] According to the temperature, humidity, wind speed, and precipitation intensity in the meteorological data, the four pollination rates at the corresponding time are obtained, the lowest value of the four pollination rates is taken, and the pollination rates per hour are accumulated to obtain the final hourly probability of pollination during the flowering period;

[0054] Pollination rate = the final hour probability of pollination during flowering period / (number of days of crop flowering period*24 hours);

[0055] If the pollen amount of crops is set to be three times the required amount, then 1 / 3 of the pollen is enough for 100% pollination of the crops;

[0056] The adjusted pollination rates are:

[0057] Pollination rate = the final hour probability of pollination during flowering period / (number of days of crop flowering period * 24 hours) / (1 / 3);

[0058] The pollination rate is multiplied by the water and fertilizer demand calculation results and the estimated yield calculation results output by the Wofost model to obtain the revised water and fertilizer demand calculation results and the estimated yield calculation results.

[0059] In this embodiment, self-pollination refers to the pollen of the stamen falling on the stigma of the pistil in the same flower. Most crops are self-pollinating, such as wheat and rice. Alienated pollination: stamens and pistils are in different flowers, common crops such as corn and fruits; effective pollination rate: that is, the pollination percentage, an indicator of the pollination ratio.

[0060] According to the above embodiment of the present invention, preferably, step S2 specifically includes: taking the sand, loam, clay, bulk density, and organic matter content data in the soil as basic data, combining the soil texture data, texture content data, time series data, and precipitation data in the meteorological data in the data set, according to the VanGenuchten water characteristic curve equation, combined with the VanGenuchten water characteristic curve related parameters for calculation. The soil water content data of the final calculation result includes soil saturated water content, field water holding capacity, and wilting point. It will be used as the basic parameter data for water calculation in the Wofost model and participate in the calculation.

[0061] In this embodiment, according to the soil survey data, the soil layer is divided into 8 depths: 4.5 cm, 9.1 cm, 16.6 cm, 28.9 cm, 49.3 cm, 82.9 cm, 138.3 cm, and 229.6 cm, and the soil water content data is calculated for each depth of the soil layer. Sand, loam, clay, bulk density, organic matter content, soil texture data, soil texture content data in the soil: source is "China Land Surface Simulation Soil Property Dataset"; precipitation data in meteorological data: source and S4, record meteorological data, obtained from NASA and the European Meteorological Center, used to calculate the soil compaction in the soil texture (the compactness of plowed soil increases after precipitation); Van Genuchten water characteristic curve: Van Genuchten model is an empirical formula that describes the soil moisture characteristic curve. Detailed public information can be found in public channels, and this application will not repeat it.

[0062] According to the above embodiment of the present invention, preferably, step S4 specifically includes: according to the effective nitrogen, phosphorus, potassium, total nitrogen, total phosphorus, total potassium, organic matter content in the soil, and the fertilization data in the agricultural data, the soil nitrogen content, phosphorus content, potassium content, mineralization data, and fertilizer loss data are calculated as inputs of the Wofost model. Effective nitrogen, phosphorus, potassium, total nitrogen, total phosphorus, total potassium, and organic matter content: the source is the "Chinese Land Surface Simulation Soil Property Dataset"; agricultural data include: fertilization date, fertilization amount, and fertilization type.

[0063] In this embodiment, according to soil survey data, the soil layer is divided into 8 depths: 4.5 cm, 9.1 cm, 16.6 cm, 28.9 cm, 49.3 cm, 82.9 cm, 138.3 cm, and 229.6 cm, and soil fertility data is calculated for each depth of the soil layer.

[0064] The specific implementation process includes:

[0065] First, establish a basic information table for green manure and chemical fertilizers to clarify the content of nitrogen, phosphorus and potassium in different fertilizers, such as shown in Table 1 and Table 2:

[0066] Table 1 Nitrogen, phosphorus and potassium content of green manure

[0067]

[0068] Table 2 shows the nitrogen, phosphorus and potassium content of fertilizers:

[0069] Table 2 Nitrogen, phosphorus and potassium content of fertilizers

[0070]

[0071] Calculate the fertility of green manure, total fertility of green manure = the sum of (green manure * (NPK content) * annual release percentage) over many years.

[0072] Daily mineralization decomposition = (total fertility of green manure + basic residue (tn, tp, tk) * 0.5) / annual average number of days above zero degrees;

[0073] Total soil fertility = daily mineralization decomposition * daily mineralization accumulation from the last planting season to now + current inorganic nutrients AN / AP / AK in the soil;

[0074] Soil fertility data will be used as basic soil fertility parameters and will be included in crop growth calculations and fertilizer demand calculations.

[0075] By modifying the irrigation date, irrigation amount, fertilization date, fertilization amount or fertilization type for multiple times, the Wofost model outputs multiple sets of crop growth status data and yield data to determine the growth status and yield of crops under different conditions, thereby obtaining the best planting date and the best harvest date.

[0076] The training process of the Wofost model further includes:

[0077] The planting management data and yield data from multiple locations and multiple growing seasons are collected in a mixed manner, and satellite remote sensing data within the growing season is obtained at the same time. The remote sensing data is coupled with meteorological data, soil water content data, farming data, and soil fertility data, and the Wofost model is obtained through deep learning.

[0078] The estimated production calculation process data and the estimated production calculation results of the Wofost model are stored and / or read on a daily basis to support breakpoint insertion of remote sensing data and API data and perform breakpoint calculations.

[0079] like Figure 2 As shown, according to another embodiment of the present invention, a water and fertilizer demand prediction system 200 for improving crop yield is also disclosed, including: a memory 201, used to store program instructions; a processor 202, used to call the program instructions stored in the memory to implement the water and fertilizer demand prediction method for improving crop yield as in the above embodiment.

[0080] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for predicting water and fertilizer demand for increasing crop yields, characterized in that: include: S1, obtaining meteorological data, the meteorological data including temperature, wind speed, evaporation, precipitation and solar radiation, and inputting the obtained meteorological data into the Wofost model; S2, calculating soil water content data, the soil water content data including: saturated water content, field water holding capacity, and wilting point, the meteorological data is used to calculate the soil water content data; S3, recording agricultural data, the agricultural data including: farmland area, planting variety, planting date, planting amount, irrigation date, irrigation amount, fertilization date, fertilization amount, and fertilization type; S4, calculating soil fertility data, the soil fertility data including: nitrogen content, phosphorus content, potassium content, mineralization data, fertilizer loss data, the farming data is used to calculate the soil fertility data; S5, setting a root depth calculation module and a pollination rate calculation module in the Wofost model, and optimizing the Wofost model according to the calculation results of the root depth calculation module and the pollination rate calculation module, wherein the root depth calculation module is used to calculate the actual root depth data of the crop, and selects the soil water content data and soil fertility data corresponding to the soil layer touched by the roots according to the actual root depth data to participate in the calculation of the Wofost model, so as to obtain the water and fertilizer demand calculation results and the estimated yield calculation results, and the pollination rate calculation module is used to calculate the pollination rate of the crop, and correct the water and fertilizer demand calculation results output by the Wofost model according to the pollination rate, and correct the estimated yield calculation results according to the pollination rate; S6, based on the optimized Wofost model, input the meteorological data, the soil water content data, the soil fertility data, and the farming data to perform yield estimation calculation, and select the yield estimation calculation result with the highest yield and the corresponding water demand data, nutrient demand data, and crop transpiration data, wherein the nutrient demand data is the nutrient demand calculated by the Wofost model in an unlimited water mode; S7, calculating daily irrigation demand according to the meteorological data, the soil water content data, the farming data and the crop transpiration data; S8, calculating daily fertilization requirements according to the soil fertility data, the farming data and the nutritional requirement data.

2. The method for predicting water and fertilizer demand for increasing crop yield according to claim 1, characterized in that: Also includes: By modifying the irrigation date, irrigation amount, fertilization date, fertilization amount or fertilization type for multiple times, the Wofost model outputs multiple sets of crop growth status data and yield data to determine the growth status and yield of crops under different conditions, thereby obtaining the best planting date and the best harvest date.

3. The method for predicting water and fertilizer demand for increasing crop yield according to claim 1, characterized in that: The S2 specifically includes: According to the soil survey data, the soil layer is divided into 8 depths: 4.5 cm, 9.1 cm, 16.6 cm, 28.9 cm, 49.3 cm, 82.9 cm, 138.3 cm and 229.6 cm. The sand, loam, clay, bulk density and organic matter content data in the soil are used as the basic data. The soil texture data, texture content data, time series data and precipitation in the meteorological data are substituted into the Van Genuchten water characteristic curve equation. The soil saturated water content, field water holding capacity and wilting point in the soil water content data are calculated in combination with the Van Genuchten water characteristic curve parameters. The soil water content data are calculated for each depth of the soil layer.

4. The method for predicting water and fertilizer demand for increasing crop yield according to claim 1, characterized in that: The S4 specifically includes: According to the soil survey data, the soil layer is divided into 8 depths: 4.5 cm, 9.1 cm, 16.6 cm, 28.9 cm, 49.3 cm, 82.9 cm, 138.3 cm and 229.6 cm. The nitrogen content, phosphorus content, potassium content, mineralization data and fertilizer loss data are calculated based on the effective nitrogen, phosphorus and potassium, total nitrogen, total phosphorus, total potassium and organic matter contents in the soil as well as the fertilization date, fertilizer amount and fertilizer type in the agricultural data. The soil fertility data are calculated for each depth of the soil layer.

5. The method for predicting water and fertilizer demand for increasing crop yield according to claim 1, characterized in that: The calculation process of the root depth calculation module specifically includes: Establish effective root depth function and root density function; According to the accumulated temperature and the root nutrient distribution function during the growth stage, the daily root dry matter accumulation and daily effective root depth were calculated; According to the daily effective root depth calculation result, the actual depth reached by the roots is determined to obtain the actual root depth data, and the soil water content data and soil fertility data corresponding to the soil layer touched by the roots are selected according to the actual root depth data as the input data of the Wofost model.

6. The method for predicting water and fertilizer demand for increasing crop yield according to claim 1, characterized in that: The calculation process of the pollination rate calculation module specifically includes: Crop pollination is classified according to self-pollination and alienation pollination, and different pollination rates are preset for different temperatures during pollination, different pollination rates are preset for different humidity during pollination, different pollination rates are preset for different wind speeds during pollination, and different pollination rates are preset for different precipitation intensities during pollination; According to the temperature, humidity, wind speed, and precipitation intensity in the meteorological data, four pollination rates at corresponding times are obtained, the lowest value among the four pollination rates is taken, and the pollination rates per hour are accumulated to obtain the final hourly probability and sum of pollination during the flowering period; Pollination rate = the final hour probability of pollination during flowering period / (number of days of crop flowering period*24 hours); If the pollen amount of crops is set to be three times the required amount, then 1 / 3 of the pollen is enough for 100% pollination of the crops; The adjusted pollination rates are: Pollination rate = the final hour probability of pollination during flowering period / (number of days of crop flowering period * 24 hours) / (1 / 3); The pollination rate is multiplied by the water and fertilizer demand calculation result and the estimated yield calculation result output by the Wofost model to obtain a revised water and fertilizer demand calculation result and the estimated yield calculation result.

7. The method for predicting water and fertilizer demand for increasing crop yield according to claim 1, characterized in that: The training process of the Wofost model includes: The planting management data and yield data of multiple locations and multiple growing seasons are collected in a mixed manner, and the satellite remote sensing data within the growing season is obtained at the same time. The remote sensing data is coupled with the meteorological data, the soil water content data, the farming data, and the soil fertility data, and the Wofost model is obtained through deep learning.

8. The method for predicting water and fertilizer demand for increasing crop yield according to claim 7, characterized in that: Also includes: The estimated production calculation process data and the estimated production calculation results of the Wofost model are stored and / or read on a daily basis to support breakpoint insertion of the remote sensing data and API data and perform breakpoint calculation.

9. A water and fertilizer demand prediction system for increasing crop yields, characterized in that: include: A memory for storing program instructions; A processor is used to call the program instructions stored in the memory to implement the water and fertilizer demand prediction method for improving crop yield as described in any one of claims 1 to 8.

Images