Method for predicting crop irrigation amount in arid saline area and related device

Abstract

The application discloses a crop irrigation amount prediction method in arid and saline-alkali regions and related equipment. The method comprises the following steps: calculating the potential evapotranspiration of the target crop based on the predicted key meteorological parameters of the target region through a dual-source evapotranspiration mechanism model, and dynamically updating the leaf area index and plant height according to the accumulated temperature process; constructing different stress inhibition functions for the concerned salt ions and the pH value in the soil solution of the saline-alkali soil, respectively, applying a composite inhibition at the stomatal conductance and / or apparent evapotranspiration resistance site to correct the potential evapotranspiration, and obtaining the ideal crop evapotranspiration corrected by the salt and pH value; and predicting the daily irrigation amount and water distribution schedule suggestion within a future preset period based on the corrected ideal crop evapotranspiration. The method can solve the problem that the current crop irrigation in arid inland saline-alkali areas relies on experience and lacks quantitative analysis of the influence caused by different salt types and contents.

Description

TECHNICAL FIELD

[0001] The embodiment of the present application relates to the field of computers, and more particularly, to a crop irrigation amount prediction method in a drought and saline area and related equipment. BACKGROUND

[0002] Accurate prediction of crop water consumption is an important prerequisite for regional water resource allocation, optimization of irrigation system, and even development of water-saving agriculture. However, due to the influence of complex environmental changes and variation of crop hydraulic characteristics, there are still many challenges in estimating crop water consumption in saline-alkali soil.

[0003] The total area of saline-alkali soil in China is about 1.5 billion mu, and the area that can be managed and utilized is about 550 million mu, which is distributed in five regions, including the coastal area, the Songnen Plain in Northeast China, the Huang-Huai-Hai Plain, Xinjiang, and the Hexi Corridor. The total salt content in Xinjiang is 5-20 grams per kilogram, and the salt mainly includes sodium chloride, sodium sulfate, and sodium carbonate, among which sodium ions, chloride ions, and sulfate ions are dominant. Salinization can lead to soil compaction and nutrient imbalance, and seriously affect crop yield. At present, through the comprehensive improvement of salt-tolerant varieties, salt washing by irrigation and drainage, chemical improvement, biological remediation, and agronomic measures, more than 30 million mu of saline-alkali soil has been restored to arable land, and there is still great potential in the future. In addition, there is a shortage of fresh water resources and unreasonable use of water resources in the current saline-alkali soil management, and there is a certain waste problem. It is necessary to calculate evapotranspiration ET to reasonably guide irrigation work. ET refers to the sum of soil evaporation and plant transpiration, which is the core link of regional water and heat balance and carbon cycle. ET estimation methods can be divided into four categories: in the micrometeorological method, the eddy correlation directly calculates ET with high-frequency three-dimensional wind speed, temperature and humidity, and carbon dioxide flux, which has high precision but expensive instruments, and is mainly used in flux stations; the Bowen ratio-energy balance method uses the energy balance equation, which is simple in equipment and suitable for field, but its coverage is limited. The water balance method calculates ET by precipitation, irrigation, seepage, and soil water storage changes, which is intuitive in principle and suitable for irrigation areas and basins, and requires high-precision soil moisture monitoring assistance; the crop coefficient method divides ET into potential evaporation and crop coefficient Kc, and combines soil moisture correction Ks, where, , which is widely used in farmland irrigation design. Among them, the crop coefficient method is the most widely used method, which usually uses the Penman-Monteith equation to calculate , but the physical principle of this method is based on the large leaf model, but due to the assumption of complete ground coverage and the problem of single stress state, it brings great uncertainty to the simulation of crop water consumption in saline-alkali soil. The Jarvis model considers the influence of meteorological factors such as VPD and air temperature on ET, but does not consider the inhibitory effect of salt on ET. SUMMARY

[0004] A series of simplified concepts are introduced in the summary section, which will be further detailed in the detailed description section. The summary section of the present application does not mean to attempt to define the key features and essential technical features of the claimed technical solutions, nor to attempt to determine the protection scope of the claimed technical solutions.

[0005] In order to solve the problem that the current crop irrigation in the saline-alkali land of arid inland area relies on experience and lacks quantitative analysis of the influence of different salt types and contents, in the first aspect, the present application provides a crop irrigation amount prediction method in a saline-alkali land area, the method comprising:

[0006] Based on the predicted key meteorological parameters of the target area, the potential evapotranspiration of the target crop is calculated by a dual-source evapotranspiration mechanism model, and the leaf area index and plant height are dynamically updated according to the accumulated temperature process;

[0007] Different stress inhibition functions are constructed for the concerned salt ions and the pH in the soil solution of the saline-alkali land, and a composite inhibition is applied at the stomatal conductance and / or apparent evapotranspiration resistance site to correct the potential evapotranspiration, so as to obtain an ideal crop evapotranspiration corrected by salt and pH;

[0008] Based on the corrected ideal crop evapotranspiration, the daily irrigation amount and water distribution schedule in the future preset period are predicted.

[0009] Optionally, it further comprises:

[0010] Based on the historical meteorological data of the target area, a multi-member rolling weather forecast is performed by a NeuralGCM weather model, and the historical meteorological data includes an ERA5 reanalysis data set.

[0011] Optionally, it further comprises:

[0012] The prediction results of the NeuralGCM weather model are taken as inputs to reforecast the key meteorological parameters by coupling a binary enzyme action optimization algorithm with a Gaussian process regression algorithm, and the key meteorological parameters are meteorological variables required by the evapotranspiration model and meteorological variables required by the leaf area index and plant height model.

[0013] Optionally, the stress inhibition function comprises:

[0014] At least two of a threshold-slope type inhibition function for sodium ions, a double-threshold interval inhibition function for chloride ions, an optimal-toxic concentration window inhibition function for sulfate ions, and a nonlinear inhibition function for pH.

[0015] Optionally, the composite inhibition applied at the stomatal conductance and / or apparent evapotranspiration resistance site to correct the potential evapotranspiration to obtain the ideal crop evapotranspiration corrected by salt and pH comprises:

[0016] The stress inhibition functions are combined in parallel or series at the stomatal conductance and / or apparent evapotranspiration resistance site to modify the potential evapotranspiration to obtain the ideal crop evapotranspiration modified by salinity and pH.

[0017] Optionally, further comprising:

[0018] The environmental stress is separated from the ionic stress using an improved Jarvis model to modify the ideal crop evapotranspiration by environmental stress.

[0019] Optionally, further comprising:

[0020] The modified ideal crop evapotranspiration is evaluated on a daily scale using Bowen ratio energy balance or eddy covariance flux observations.

[0021] In a second aspect, the present application further provides a crop irrigation amount prediction device for a drought and saline-alkali region, comprising:

[0022] A calculation unit is configured to calculate a potential evapotranspiration of a target crop based on predicted key meteorological parameters of a target region by a dual-source evapotranspiration mechanism model, and dynamically update a leaf area index and a plant height according to an accumulated temperature process;

[0023] A modification unit is configured to construct different stress inhibition functions for a concerned salt ion and pH in a soil solution of a saline-alkali land, respectively, modify the potential evapotranspiration by applying a composite inhibition at a stomatal conductance and / or apparent evapotranspiration resistance site, and obtain an ideal crop evapotranspiration modified by salinity and pH.

[0024] A prediction unit is configured to predict a daily irrigation amount and a water distribution schedule suggestion in a future preset period based on the modified ideal crop evapotranspiration.

[0025] In a third aspect, an electronic device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor is configured to implement the steps of the crop irrigation amount prediction method for a drought and saline-alkali region according to any one of the first aspect.

[0026] In a fourth aspect, the present application further provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executable by a processor to implement the crop irrigation amount prediction method for a drought and saline-alkali region according to any one of the first aspect.

[0027] In summary, the crop irrigation amount prediction method in a drought and saline area provided in the present application calculates the potential evapotranspiration of the target crop through the double-source evapotranspiration mechanism model based on the predicted key meteorological parameters of the target area, and dynamically updates the leaf area index and plant height according to the accumulated temperature process; different stress inhibition functions are respectively constructed for the concerned salt ions and the pH value in the soil solution of the saline land, the composite inhibition is applied at the stomatal conductance and / or apparent evapotranspiration resistance site, the potential evapotranspiration is corrected, and the ideal crop evapotranspiration corrected by the salt and pH value is obtained; based on the corrected ideal crop evapotranspiration, the daily irrigation amount and water distribution schedule in the future preset period are predicted. Through the ion and pH stress function of different categories and different intervals, and by applying at the clear mechanism site, the different physiological effects of sodium ions and sulfate ions under the same EC can be quantitatively presented, and different water distribution strategies and water amount intervals are directly generated. The LAI or H dynamic update driven by the accumulated temperature ensures that the resistance network and energy distribution conform to the current community structure, avoids misjudgment of the salt effect due to the difference in growth period, and thus improves the credibility of the water requirement estimation. The corrected evapotranspiration sequence can be converted into the daily recommended water amount and shift schedule, the canal system efficiency, the leakage upper limit and the target water content are considered, and the water saving target and the field pipe water constraint are compatible.

[0028] The drought and saline area crop irrigation amount prediction method of the present application, other advantages, objects and features of the present application will be embodied in part by the following description, and will be understood by those skilled in the art through the study and practice of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0029] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are for purposes of illustration only and are not considered a limitation of the present specification. Moreover, like reference numerals are used to designate like parts throughout the specification and drawings. In the drawings:

[0030] Figure 1 A drought and saline area crop irrigation amount prediction method flowchart provided by an embodiment of the present application;

[0031] Figure 2 A binary encoding diagram of 50 members in a drought and saline area crop irrigation amount prediction method provided by an embodiment of the present application;

[0032] Figure 3 A schematic flowchart of a binary enzyme action optimization algorithm coupled with a Gaussian process regression for daily net radiation prediction in a drought and saline area crop irrigation amount prediction method provided by an embodiment of the present application;

[0033] Figure 4 A growth function diagram of plant height and leaf area index in an embodiment of the present application;

[0034] Figure 5 It is the forecast value comparison diagram of the highest temperature observation value, different forecast period forecast value and comparison method in the embodiment of the application.

[0035] Figure 6 It is the forecast value comparison diagram of the lowest temperature observation value, different forecast period forecast value and comparison method in the embodiment of the application.

[0036] Figure 7 It is the forecast value comparison diagram of the relative humidity observation value, different forecast period forecast value and comparison method in the embodiment of the application.

[0037] Figure 8 It is the forecast value comparison diagram of the wind speed observation value, different forecast period forecast value and comparison method in the embodiment of the application.

[0038] Figure 9 It is the forecast value comparison diagram of the net radiation observation value, different forecast period forecast value and comparison method in the embodiment of the application.

[0039] Figure 10 It is the forecast value comparison diagram of the ET observation value, different forecast period forecast value and comparison method in the embodiment of the application.

[0040] Figure 11 It is a structure schematic diagram of a crop irrigation amount prediction device provided by the embodiment of the application in a drought and saline area.

[0041] Figure 12 It is a structure schematic diagram of a crop irrigation amount prediction electronic device provided by the embodiment of the application in a drought and saline area. DETAILED DESCRIPTION

[0042] In order to solve the problem that current crop irrigation in a drought and saline area is based on experience and lacks quantitative analysis of the influence of different salt types and contents, please refer to Figure 1 It is a flowchart of a crop irrigation amount prediction method provided by the embodiment of the application in a drought and saline area, which can specifically include steps S110 to S130.

[0043] S110, based on the predicted key meteorological parameters of the target area, the potential evapotranspiration of the target crop is calculated by a dual-source evapotranspiration mechanism model, and the leaf area index and plant height are dynamically updated according to the accumulated temperature process.

[0044] S120, different stress inhibition functions are constructed for the concerned salt ions and the pH value in the soil solution of the saline land, the ideal crop evapotranspiration amount is corrected by applying a composite inhibition at the stomatal conductance and / or apparent evapotranspiration resistance site, and the ideal crop evapotranspiration amount corrected by the salt and pH value is obtained.

[0045] S130, predicting daily irrigation amount and water distribution schedule in a future preset period based on the corrected ideal crop evapotranspiration.

[0046] It can be understood that in the arid inland saline-alkali land, the actual evapotranspiration of crops is not only controlled by meteorological conditions and community structure such as leaf area index LAI and plant height H, but also affected by physiological stress of soil solution ion composition and pH. Then the dual-source evapotranspiration mechanism model can be used to physically decompose the canopy transpiration and soil evaporation to obtain the potential evapotranspiration without salt-alkali inhibition, and then dynamically update LAI and H according to the accumulated temperature process, so that the energy exchange and aerodynamic resistance always match the current growth state. On this structural foundation, stress inhibition functions are established for sodium ion, chloride ion, sulfate ion and pH respectively, and the force application site is determined at stomatal conductance and / or apparent evapotranspiration resistance, so that the functional constraint of ions and pH on the transpiration channel can be quantitatively presented, thereby obtaining the ideal crop evapotranspiration corrected by salt and pH. Finally, taking the corrected daily evapotranspiration as the water demand baseline, combined with the effective soil layer depth, target water content, canal / drip irrigation efficiency and leakage rate, the daily irrigation amount and executable water distribution schedule in a future preset period are derived. The pain point of being unable to distinguish the influence of different salt types and contents by experience alone is solved.

[0047] For example, the dual-source evapotranspiration model is driven by key meteorological parameters such as daily net radiation, air temperature, relative humidity and near-ground wind speed to decompose the surface flux into canopy transpiration and soil evaporation. The model uses energy conservation and resistance network as the skeleton on a daily scale, and uses aerodynamic resistance, canopy resistance and soil surface resistance to construct the coupling solution of sensible heat and latent heat terms; the canopy coverage coefficient is determined by LAI, and the roughness and zero plane displacement are determined by plant height H. In order to avoid systematic distortion caused by fixed structure, LAI and H are not constant, but are updated daily by GDD (Growing Degree Days, crop accumulated temperature), using a continuous function family such as logarithmic-logistic or piecewise smooth, whose output directly affects the canopy coverage, aerodynamic parameters and resistance chain, thereby determining the potential transpiration channel capacity. Because the vigorous growth period with higher LAI and H and the seedling stage with lower LAI and H have completely different energy distribution and transmission efficiency under the same meteorological forcing, if not dynamically updated, the subsequent salt-alkali stress will be wrongly attributed, resulting in deviation in water demand estimation. Therefore, without considering salt-alkali, the potential evapotranspiration sequence consistent with the current growth state is obtained, which avoids the systematic deviation caused by growth period mismatch compared with fixed Kc or fixed structure, and ensures the accuracy of the physical carrier of subsequent stress application.

[0048] Exemplarily, the soil solution chemistry of the field can be measured to select ions and pH that are of interest for the target crop and the region, such as sodium, chloride, sulfate, and pH. A single-factor inhibition function with a distinguishable threshold and a sensitive interval can be constructed for each factor, and the output ranges from 0 to 1, representing the functional inhibition intensity on stomatal conductance or apparent evapotranspiration resistance. The configuration can include: a threshold-slope function for sodium ions, i.e., weak impact below the starting threshold and rapid inhibition above the critical threshold; a double-threshold interval function for chloride ions, i.e., a segmented continuous curve of a safe zone-excessive zone-toxic zone; an optimal-toxic concentration window for sulfate ions, i.e., both too low and too high are not conducive, and there is an optimal window; a nonlinear function for pH, i.e., the farther the deviation from the suitable pH of the crop, the stronger the inhibition. To avoid confusion with environmental stress, these salt and pH inhibitions are explicitly applied to the stomatal conductance and / or apparent evapotranspiration resistance site, which can be equivalent to multiplying the stomatal conductance by a comprehensive inhibition coefficient or dividing the canopy resistance by a comprehensive inhibition coefficient in the canopy channel, and introducing a corresponding inhibition factor in the soil surface resistance if evidence shows that high salt leads to a decrease in surface osmotic potential and a limitation of capillary recharge. The combination of multiple inhibitions can use a parallel product or a coupled weight average, and the synthesis method that can better explain the observations can be selected through a validation set. In this way, the effects of different ion species and different concentrations, different pH deviations on the transpiration channel are quantitatively distinguished and decoupled from environmental stress in the same physical channel, and an ideal crop evapotranspiration that can explain and reproduce experimental trends is obtained. The results can directly support the fact that the same EC is not equivalent, but the composition, content, and pH jointly determine the water demand difference.

[0049] Exemplarily, the modified daily evapotranspiration sequence can be used as a water demand baseline to calculate the replenishment amount and form a shift scheduling water schedule according to the effective soil layer depth of the root layer, the current measured water content, the target water content, and the canal / drip irrigation efficiency parameters. In this way, the target water gap and the recent water demand are used together to determine the recommended water amount for each day, and the upper limit can be subject to the leakage rate and the field water holding capacity. The scientific quantification of ideal crop evapotranspiration is converted into an executable schedule, while explicitly considering canal efficiency and leakage upper limit, resulting in a water-saving and controllable risk water allocation plan.

[0050] According to some embodiments, further comprising:

[0051] Based on historical meteorological data of the target area, a multi-member rolling weather forecast is made by a NeuralGCM weather model, and the historical meteorological data includes an ERA5 reanalysis dataset.

[0052] The forecast results of the NeuralGCM weather model are input to a Gaussian process regression algorithm driven by a binary enzyme action optimization algorithm to reforecast key meteorological parameters required by the evapotranspiration model.

[0053] Exemplary, with ERA5 reanalysis as meteorological state and boundary history to recent constraints, drive a neural network type global weather model NeuralGCM to generate memberized and time-sliced rolling forecast sequences in target region, ensemble members characterize initial value and model uncertainty, rolling manner ensures continuous coverage of future period for any starting date. Can obtain ERA5 ground and multi-isobaric surface elements within target region neighborhood, unified to model required time step and grid, with recent several days of ERA5 assimilation window as starting state, trigger NeuralGCM to generate an ensemble of no less than 50 members, forecast period covers 1-15 days, time slicing contains at least daily scale, unify output fields in height and unit, such as wind speed converted to 2m, net radiation converted to daily cumulative, form a candidate set of original driving required by evapotranspiration model. Compared with single-path numerical prediction, ensemble form provides prediction mean and dispersion, which is convenient for subsequent transmission of meteorological uncertainty to the safety boundary of irrigation interval.

[0054] Exemplary, directly using ensemble output to drive mechanism model will bring in systematic bias and time-spread. Therefore, a binary enzyme action optimization can be used to make 0 / 1 selection in high-dimensional, multi-member, multi-level candidate features, to obtain a subset of key meteorological parameters such as daily net radiation, maximum / minimum temperature, relative humidity and near-surface wind speed that best explain the features. Then, the selected features are used as input to train a Gaussian process regression to re-predict these key parameters at site or field scale, and modeling is performed for different forecast periods. This can significantly reduce the systematic bias of the original ensemble output, make the driving quantity consistent with the measured station / field flux station scale, and improve the stability and credibility of evapotranspiration and irrigation quantity recommendations under medium and long-term time.

[0055] According to some embodiments, the stress inhibition function comprises:

[0056] At least two of a threshold-slope type inhibition function for sodium ions, a double-threshold interval inhibition function for chloride ions, an optimal-toxic concentration window inhibition function for sulfate ions, and a nonlinear inhibition function for pH.

[0057] It can be understood that thus, even if EC is similar, the difference in ion inhibition will bring different transpiration capacity and water distribution strategies, which can solve the misdiagnosis problem of the same EC and the same strategy. The over-threshold steep drop of sodium ions, the safe-to-excessive-toxic of chloride ions, and the optimal window of sulfate ions are all parameterized and parameterized, which can be corrected according to experimental or historical data, and have portability. The function does not directly modify LAI or H, but maps to functional channels such as canopy resistance or stomatal conductance, ensuring consistency with the energy-resistance structure of two-source evapotranspiration, and reducing misjudgment. The daily water quantity and shift generated directly from ideal ET, and the prediction uncertainty can be mapped to the safety boundary, which is convenient for balancing between water saving and stable production.

[0058] According to some embodiments, the applying compound inhibition at the stomatal conductance and / or apparent evapotranspiration resistance site modifies the potential evapotranspiration amount to obtain a salt and pH corrected ideal crop evapotranspiration amount, including:

[0059] The stress inhibition function is combined in parallel or in series with a compound inhibition coefficient at the stomatal conductance and / or apparent evapotranspiration resistance site to modify the potential evapotranspiration amount to obtain a salt and pH corrected ideal crop evapotranspiration amount.

[0060] It is necessary to explain that in the model layer, the dual-source evapotranspiration mechanism of canopy and soil flux decomposition is adopted, the energy input is distributed to the canopy and soil according to the coverage, and the network of aerodynamic resistance, canopy resistance and soil surface resistance is used to solve the daily scale potential transpiration and evaporation; In order to avoid the system error caused by static structure, the leaf area index and plant height are updated day by day according to the accumulated temperature process, so that the key structural quantities such as coverage, roughness and zero plane displacement are consistent with the growth process. In actual implementation, the initial value is given by historical observation or variety file, and then the accumulated temperature threshold or smooth growth function is used to automatically advance, if necessary, supplemented by remote sensing or sample strip investigation for correction. The potential evapotranspiration and channel allocation consistent with the current growth state can be obtained, which provides an accurate structural foundation for the subsequent physical application of stress, and avoids misjudging the difference between growth periods as salt influence. For the concerned factors in soil solution, functions are established for sodium ion, chloride ion, sulfate ion and pH respectively, each function output is limited between zero and one to represent the continuous degree of no to strong inhibition, in implementation, first collect the control variable test or historical field data, calculate the baseline transpiration or stomatal conductance under low salt or suitable pH conditions, then use the relative changes under different treatments to fit the shape parameters and threshold parameters of each function, after fitting, cross validation is carried out to check that the low risk area should be close to no inhibition, over threshold or deviate from the optimal window should decrease rapidly, if necessary, establish parameter family at crop variety or soil texture level to improve generalization. The physiological effects of ion species and content, pH deviation are converted into quantitative functions that can be directly called to form a portable and reproducible inhibition module. In site selection, the inhibition function is applied to the canopy stomatal conductance or equivalent canopy apparent resistance first to express that ions and pH mainly regulate the main channel of transpiration through stomata, when monitoring shows that salt crust or high osmotic pressure leads to soil evaporation limited, then introduce soil side inhibition in soil surface resistance site; In synthesis, if each factor can act independently, parallel combination is adopted, the single factor inhibition is multiplied to form a canopy composite inhibition coefficient with equal weight or verified weight, and the stomatal conductance is scaled or the apparent resistance is enlarged accordingly; If there is a clear physiological sequence, such as osmotic stress caused by sodium ion first, and enzyme inactivation aggravated by alkaline, then adopt series combination, update the stomatal conductance in the specified order one by one, so that the post factor continues to exert force on the limited channel, and the lower limit of the inhibition coefficient should be set to avoid numerical rigidity, and the parallel and series combination should be objectively selected through fitting degree and diagnostic variables such as leaf temperature. In this way, accurate grasp of the mechanism difference of multiple factors acting together can not only express the short board effect, but also express the hierarchical pressure, so as to improve the authenticity and stability of stress representation. After completing the site and synthesis strategy, the canopy composite inhibition coefficient is applied to the stomatal conductance or canopy apparent resistance, and if necessary, the soil side inhibition is applied to the soil surface resistance, and then the daily scale canopy transpiration and soil evaporation are re-solved in the same energy-resistance framework to obtain the ideal crop evapotranspiration modified by salt and pH.To ensure physical consistency, the actual evapotranspiration should not exceed the potential evapotranspiration, the canopy and soil components should not be negative, and other constraints are set during the solving process. The upper limit of resistance and iteration relaxation are used to improve the convergence in extreme inhibition or extreme weather conditions. The chemical to physiological inhibition is seamlessly embedded in the physical to energy flux calculation, and the ideal evapotranspiration time sequence consistent with the component distinguishable and the growth period structure is output.

[0061] According to some embodiments, further comprising:

[0062] The modified ideal crop evapotranspiration is evaluated on a daily scale using Bowen ratio energy balance or eddy covariance flux observations.

[0063] It can be understood that a flux observation station can be arranged in the target field, and sensible heat, latent heat and energy items can be obtained by using Bowen ratio energy balance or eddy correlation method, and observed evapotranspiration can be obtained by aggregating on a daily scale to evaluate the ideal crop evapotranspiration modified by salt and pH. Thus, a daily scale reality check on the same scale as the mechanism network can be provided to ensure that the effects of ion composition and pH modification are derived from functional sites rather than driving noise. The robustness in extreme hot and dry and high salt scenarios can be improved through hierarchical and closed correction, so that the water saving and water distribution recommendations have a quantifiable and reliable interval. And it can form a self-calibration channel for long-term operation, so that the parameters remain traceable and consistent with the year and management conditions.

[0064] It can be understood that the above method is applicable to saline-alkali land in the inland arid area of Xinjiang, and other soda saline-alkali land is not suitable for the present application. In addition, when the groundwater level is within 1m, the above method is not applicable, and the underground drainage should be considered to reduce the groundwater level first. According to some embodiments, the crop irrigation amount prediction method in the above-mentioned arid saline-alkali region can comprise:

[0065] S1, the historical meteorological information of the target region for more than 3 years and the meteorological data of ERA5 reanalysis dataset can be obtained. In this step, late-maturing cotton varieties can be selected as irrigation objects, and 2-year tests can be carried out, wherein the data in 2022 can be used for model parameter calibration of the above-mentioned prediction method and comparative method, and the data in 2023 can be used for model verification. The test can be carried out in an area of 120m 2on the plot. The planting pattern can be one film 3 pipe 6 rows, and the cotton distance is 10-66-10-66-10 cm, the film spacing is 40 cm, and the plant spacing is 12 cm. The obtained meteorological data is the meteorological data of the target area weather station in 2022-2023, including daily scale 1.5 m and 2.9 m high temperature, 1.5 m and 2.9 m high wind speed, relative humidity and net radiation. The ERA5 reanalysis data can include daily earth surface sea ice coverage ratio and sea surface temperature, and 12-hour scale 37 pressure layers (1, 2, 3, 5, 7, 10, 20, 30, 50, 70, 100, 125, 150, 175, 200, 225, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 775, 800, 825, 850, 875, 900, 925, 950, 975, 1000) cloud internal specific liquid water content, cloud internal specific solid water content, air temperature, normal wind speed, zonal wind speed and potential height. The 0.25° spatial point information is interpolated to 1.4° spatial resolution grid points using bilinear interpolation method:

[0066]

[0067] The 1.4° spatial resolution correction value is obtained by using area weighting again:

[0068] .

[0069] Wherein, (x0, y0), (x0, y1), (x1, y0) and (x1, y1) are the latitude and longitude (°) of the four surrounding grid points, 11 , Q 21 , Q 12 , Q 22 are the weight coefficients of the four grid points, q i,j is the meteorological variable value, is the latitude value.

[0070] S2, the multi-member rolling weather meteorological forecast can be performed using the NeuralGCM weather large model to generate a 50-scenario future 15-day member set. The ERA5 data in S1 can be imported into the NeuralGCM weather large model to perform future 15-day rolling prediction. The 50-scenario future 15-day member set data includes 50-member information of 500 hPa, 850 hPa and 1000 hPa air temperature, specific humidity, zonal wind speed, normal wind speed and potential height every 12 hours from 12 to 360 hours.

[0071] S3, as Figure 3As shown, the prediction results can be taken as input to predict the daily total surface radiation, maximum and minimum temperature, relative humidity and 10m high wind speed for the next 15 days, respectively. When predicting the total surface radiation on the first day, the input factors are the temperature, specific humidity, vertical wind speed, normal wind speed and potential height at different pressure layers (500 hPa, 700 hPa, 850 hPa and 1000 hPa) at the 12th hour as the model input. The NeuralGCM weather model can be coupled and driven by a binary enzyme optimization algorithm to drive the Gaussian process regression algorithm. As shown, Figure 2 The 50-member set of the next 15 days of the 50 scenarios can include 50-member information of the temperature, specific humidity, zonal wind speed, normal wind speed and potential height at 500 hPa, 850 hPa and 1000 hPa at a 12-hour scale. The binary enzyme optimization algorithm can refer to, first, selecting the best position according to the objective function in the population , determining the iteration attenuation factor LF t :

[0072]

[0073] wherein, t is the current generation number, MaxIter is the maximum number of iterations. A matrix containing only 0 and 1 with the same feature dimension as the matrix can be generated, and the number of rows of the matrix is the number of substrates in each generation. When iterating, each individual in each generation will generate two substrate candidate positions, respectively, and :

[0074]

[0075] wherein, is the first substrate candidate position in the t-th generation, is the best position of the substrate in the t-1-th generation, is a random number, and e represents element-wise multiplication. The first substrate candidate position is subjected to Sigmod transformation to become a number between 0 and 1.

[0076] According to the random number, there is a probability of changing to 0 or 1:

[0077]

[0078] Two individuals in the population can be randomly selected x and y , and the distance vector is calculated: , the second substrate candidate position in the t-th generation is determined, The second substrate candidate position is subjected to Sigmod transformation to become a number between 0 and 1:

[0079] ,

[0080] According to the time number has the probability of changing to 0 or 1,

[0081]

[0082] Compare 2 substrate candidate positions to retain the best, The optimal substrate position is compared with the position of the previous generation population, and the optimal position is updated:

[0083]

[0084] Update the optimal position of the population, The binary enzyme action optimization algorithm coupled with Gaussian process regression can improve the errors in ERA5 data and be passed to the NeuralGCM large model. The daily net radiation prediction method is as follows: for the first day, the temperature, specific humidity, and potential height at different pressure layers (500 hPa, 850 hPa, and 1000 hPa) at 12 o'clock are used as model inputs, the binary enzyme action optimization algorithm coupled with Gaussian process regression is used for model training, and the ground observed net radiation is used as the target value. The data is randomly shuffled and divided into 70% and 30% two parts, of which 70% is used for model feature selection and 30% is used for testing. For the second day, the data is replaced with the 36h member set of the prediction period as the model input, and for the third day, the data is replaced with the 60h member set of the prediction period as the model input. For the next day, increase the data by 24h and so on. The daily maximum temperature prediction method is as follows: for the first day, the temperature member set at different pressure layers at 12 o'clock is used as the model input, the binary enzyme action optimization algorithm coupled with Gaussian process regression is used for model training, and the ground observed daily maximum temperature is used as the target value. The data is randomly shuffled and divided into 70% and 30% two parts, of which 70% is used for model feature selection and 30% is used for testing. For the second day, the data is replaced with the 36h member set of the prediction period as the model input, and for the third day, the data is replaced with the 60h member set of the prediction period as the model input. For the next day, increase the data by 24h and so on. The daily minimum temperature prediction method is as follows: for the first day, the temperature member set at different pressure layers at 24 o'clock is used as the model input, the enzyme action optimization algorithm coupled with Gaussian process regression is used for model training, and the ground observed daily maximum temperature is used as the target value. The data is randomly shuffled and divided into 70% and 30% two parts, of which 70% is used for model feature selection and 30% is used for testing. For the second day, the data is replaced with the 36h member set of the prediction period as the model input, and for the third day, the data is replaced with the 60h member set of the prediction period as the model input. For the next day, increase the data by 24h and so on. The daily relative humidity prediction method is as follows: for the first day, the temperature and specific humidity member sets at different pressure layers at 12 o'clock and 24 o'clock are used as the model input, the binary enzyme action optimization algorithm coupled with Gaussian process regression is used for model training, and the ground observed 2m high relative humidity is used as the target value. The data is randomly shuffled and divided into 70% and 30% two parts, of which 70% is used for model feature selection and 30% is used for testing. For the second day, the data is replaced with the 36h member set of the prediction period as the model input, and for the third day, the data is replaced with the 60h member set of the prediction period as the model input. For the next day, increase the data by 24h and so on.The daily 2m high wind speed prediction method is as follows: for the first day, the temperature, vertical wind speed and normal wind speed member sets of different pressure layers at 12 o'clock and 24 o'clock are used as model inputs, a binary enzyme action optimization algorithm is used to couple a Gaussian process regression and ground observation of high wind speed at 10m of the day as a target value for model training, the data is randomly shuffled and divided into 70% and 30% two parts, wherein 70% is used for model feature selection and 30% is used for inspection. For the second day, the data is replaced with the 36th member set of the prediction period as the model input, for the third day, the data is replaced with the 60th member set of the prediction period as the model input, and for the following days, the data is increased by 24h every day.

[0085] S4, the current stressed and well-grown crop information and growth environment information can be obtained. Before irrigation begins, 20 cotton plants are randomly selected to measure their height, a LI-3000C type leaf area index instrument is used, more than 10 locations are selected to measure the soil moisture content in the surface layer of 0-10cm using the drying method, and the surface layer soil saturation water content is measured using the ring cutter saturation-drying method or the centrifuge drying method. The saline-alkali soil type in this region is inland arid saline-alkali soil.

[0086] S5, the crop evapotranspiration can be predicted using a dual-source model and predicted meteorological factors to obtain the crop evapotranspiration in the next 15 days.

[0087]

[0088]

[0089]

[0090] Wherein, Δ is the slope of the saturated water vapor pressure curve at the current temperature (unit: ); c p Cp is the specific heat capacity of air at constant pressure; ρ is the air density (unit: );D is the water vapor pressure deficit (unit: kPa); Rc is the canopy resistance (unit: ); Rb is the overall boundary layer resistance of the vegetation component in the canopy (unit: ); Ra is the aerodynamic resistance between the average canopy airflow and the reference height (unit: ); Rd is the aerodynamic resistance between the soil surface and the average canopy airflow (unit: ); Rs is the soil surface resistance (unit: ). γ is the humidity constant (kPa·°C -1 ); , , , , A Total energy of a complete crop (MJm) -2 ·d -1 Rn is the net radiation (MJm) -2 ·d -1 G represents soil heat flux (MJ / m³). -2 ·d -1 Since crops can only absorb a portion of the energy, the portion of energy that the canopy cannot retain is denoted as . ,in A s Energy received by the matrix (MJm) -2 ·d -1 The calculation method for water vapor pressure deficit is as follows: ,in , T represents the average temperature (°C). It cannot be obtained directly; an intermediate variable is required. Seeking, , , The evaporation resistance of the soil surface layer is reflected and obtained using empirical formulas. Parameter C is the extinction coefficient. and These are soil saturated water content and soil moisture content, respectively. and Both can be estimated using canopy cover. , , ,in and These represent the aerodynamic drag between the average canopy airflow and the reference height under dry soil surface conditions, and the aerodynamic drag between the soil surface below the dry soil surface and the average canopy airflow, respectively. and These represent the aerodynamic drag between the average canopy airflow and the reference height when LAI is 3, and the aerodynamic drag between the soil surface and the average canopy airflow when LAI is 3. The potential growth in plant height over the next 15 days should be calculated considering both crop growth and senescence.

[0091]

[0092] The potential growth of leaf area index in the next 15 days should also take into account the decline caused by leaf senescence later on.

[0093]

[0094] See also Figure 4 ,Figure 4 Figure 2 (a) is the growth process of plant height with cumulative temperature, Figure 4 Figure 2 (b) is the growth process of leaf area index with cumulative temperature, parameters a , b and c should be obtained using fitting or optimization algorithms, and their values are 1.3, 0.25 and 2.2 respectively. Parameters a s and b s should be obtained using fitting or optimization algorithms, and their values are 24 and 3.8 respectively. Other parameters are shown in Table 1.

[0095] Table 1

[0096]

[0097] S6, the salt stress impact function can be calculated. The crop is affected by salt and alkalinity stress impact functions, salt includes sodium ion, chloride ion and sulfate ion, and pH stress impact.

[0098]

[0099] wherein the sodium ion stress K sNa Impact on ET:

[0100] .

[0101] The sulfate ion stress K ss Impact on ET:

[0102]

[0103] wherein S0 and S1 are the optimum concentration and the crop is affected by the highest toxic concentration.

[0104] The chloride ion stress K sCl Impact on ET:

[0105]

[0106] wherein C1 is the critical initial inhibition concentration, and C2 is the highest toxic concentration affecting the crop.

[0107] The stress pH K pH Impact on ET:

[0108] ​​​

[0109] Among them, parameters C min , C crit , n Obtained through fitting. C min It is 75. C crit It is 168. n The value is 3.3. Parameters S 0, S 1. Obtained using fitting. S 0 is 420. S 1 is 980, parameter C 1, C 2 was obtained using fitting, where C 1 is 240, C 2 is 530. k p It is -0.21. H ET50 It is 9.1.

[0110] The improved Jarvis model can be used to consider environmental stress over the next 15 days, with the following equation:

[0111]

[0112] in, Empirical parameters a D It is 0.0613. Empirical parameters a T The value is 0.00442, and the parameter T0 is 32℃.

[0113] The experiment can use the Bowenbi weather station to observe the actual evapotranspiration.

[0114]

[0115] in, and These represent the temperature difference and water vapor pressure difference between the upper and lower levels, respectively. The Bowenbi weather station consists of a CR1000 data acquisition unit, 083E-1 temperature and humidity sensors (installed at heights of 2.9m and 1.5m respectively), an NRLite2 net radiation meter (installed at a height of 1.5m), and an HFP01 heat flux plate (buried at a depth of 0.05m).

[0116] To demonstrate the superiority of this invention, a comparison was made using the traditional crop coefficient method. The calculation formula is as follows:

[0117]

[0118] Where ET0can be estimated using the FAO56 Penman-Monteith equation:

[0119]

[0120] Where, T mean is the average air temperature, u 2 is the wind speed at 2 m height, D is the daily scale saturated pressure difference. The required meteorological variables are daily maximum and minimum temperature, 10 m height wind speed, relative humidity and net radiation. The daily maximum and minimum temperature, 10 m height wind speed, relative humidity are converted from the 1000 hPa 50-member variable mean values predicted by NeuralGCM, and the wind speed conversion , z w is the surface roughness length. The temperature conversion , is 9.8 °C km -1 , is the height difference between 1000 hPa and 2 m, and 1000 hPa is 111 m. The specific humidity q 1000 is converted to water vapor pressure e. , where P is the standard atmospheric pressure. The NeuralGCM model does not have a net radiation prediction result, and is converted using the predicted air temperature. K c The crop coefficient is taken from Table 2.

[0121] Table 2

[0122]

[0123] Salt stress K s,salt The segmented function of conductivity can be used,

[0124]

[0125] Where, EC is the electrical conductivity of saline-alkali soil, EC l is the lower limit of cotton stress, taken as 4 s·m -1 , EC u is the upper limit of cotton stress, taken as 10 s·m -1 .

[0126] Table 3 is a comparison of the statistical results of the present method and the comparative method. For example, Figures 4-10and Table 3 can be known that the method is significantly better than the comparative method in all variables and three forecast horizons. The error index shows a monotonic trend of widening gap as the forecast period extends: the root mean square error RMSE of ET decreases by 17%, 32%, and 41% at 1, 7, and 15 days, respectively; RH decreases by 26%, 32%, and 24%; Rn decreases by 28%, 30%, and 31%; Tmax decreases by 21%, 18%, and 38%; Tmin decreases by 8%, 28%, and 26%; and wind speed decreases by 15%, 34%, and 18%, showing that the long-term advantage is more obvious. Correlation increases by 5-30% on all variables, among which ET, Rn, Tmax, and Tmin increase by 15 days still remain above 0.70, while the comparative method has dropped to 0.37-0.65, indicating that the method has stronger capturing ability for nonlinear characteristics. The absolute value of pBias of all variables is <0.05, and the comparative method is up to -0.37 at 15 days, proving that the method has no systematic deviation. In summary, the method reduces the 15-day forecast error of each variable by 30-50%, which can significantly improve the reliability of irrigation forecasting.

[0127] Table 3

[0128]

[0129] It should be noted that the observed meteorological data RH is the relative humidity at 1.5 m, Tmax is the maximum temperature at 1.5 m, Tmin is the minimum temperature at 1.5 m, and Ws is the high wind speed at 1.5 m. The forecast data is at a height of 2 m, which is converted by the following formula:

[0130] For temperature: T 1.5 = T2- ΔT = T2- 0.5m × 6.5℃ / 1000m = T2- 0.0325

[0131] For wind speed: , where d = 0.7H, H is the plant height.

[0132] For relative humidity: RH 1.5 ≈ RH2.

[0133] It can be understood that wind erosion and evaporation form a high salt crust in the surface layer of 5 cm, but the salt concentration in the root activity layer of 10-30 cm is lower, and the soil evaporation is usually represented by the surface or mixed sample of the root zone, resulting in that the soil evaporation is mistakenly regarded as stomatal inhibition, and the water distribution strategy is misjudged. Considering that the salt has different channel effects on soil evaporation and stomatal conductance, the vertical salt stratification will change the relative contribution of the canopy and the soil component. In order to solve the above problems, in some examples, three-layer sampling depth can be specified and stratification factors can be established to infer the effective osmotic potential of each layer by conductivity and ion concentration; the stratification factor is written into the input dictionary. When the surface salt crust reaches the threshold and the root layer is relatively low in salt, only the soil surface resistance site is applied with inhibition, and the stomatal site is kept from being affected by the salt crust. After the net radiation is allocated according to the leaf area index, the stratification factor is adjusted preferentially to the soil component to prevent the soil side attenuation from being mistakenly attributed to the canopy. The difference between the soil surface temperature rise and the leaf temperature rise in the noon period is used for rapid diagnosis, and if the soil surface temperature rise is significantly higher than the leaf temperature, it is confirmed that the soil site inhibition is dominant. Therefore, the water distribution can be avoided to be excessive due to channel misjudgment.

[0134] It can be understood that the soil water content is sufficient in the morning, and the high vapor pressure deficit in the afternoon is superimposed with salt, resulting in stomatal closure. Even if the weather is alleviated the next day, the stomata still remain short-term memory bias closed. The conventional daily scale model ignores this effect. Considering that the stomata have short-term memory of ion and osmotic stress, which is manifested as intraday and cross-day response delay. In order to solve the above problems, in some examples, the maximum stress memory amount can be introduced, which is updated by the strongest ion inhibition and vapor pressure deficit of the previous day, and a slow release coefficient is set to decay across days. The memory amount is mapped to an additional inhibition factor at the stomatal site, which is only applied in the morning of the next day, and is slowly released in the morning according to the environment. The leaf temperature-air temperature difference exceeds the threshold and the sudden drop of carbon assimilation in the afternoon are used as trigger markers to update the memory amount. When the memory amount is high and the afternoon is predicted to be dry and hot, it is automatically suggested to advance a small irrigation class to the morning to reduce the afternoon peak. In this way, the robustness of cross-day prediction is improved, and the misjudgment caused by the extreme of the previous day is reduced.

[0135] It can be understood that strong winds bring warm and dry air advection, which makes the leaf temperature rise and the sensible heat increase. If not corrected, the advection effect is easy to be mistaken for salt inhibition. Considering that advection changes the apparent performance of energy closure and leaf stomatal behavior, advection diagnosis is needed before site inhibition. In order to solve the above problems, in some examples, an advection index can be constructed by wind speed, stability, upstream bare ground ratio and energy closure gap, and when it exceeds the threshold, the advection period is marked. During the marked period, the aerodynamic resistance is preferentially corrected in stages, and then the site inhibition is calculated; avoid mistaking the leaf temperature rise caused by advection as ion inhibition. This correction is only applied to the high wind period in the afternoon to reduce the interference with normal periods. The upstream surface type and wind corridor remote sensing layer are combined to check the advection determination. In this way, false positives are significantly reduced in windy desert boundary areas.

[0136] It can be understood that the full-area uniform drip irrigation is easy to push the salt to the periphery of the root system, and the local dryness can migrate the salt along the lateral direction under the joint action of the electric field and the water potential gradient, but no one systematically utilizes it. Considering that the mild gradient of the root zone water potential is generated in the space under the condition that the total water amount is kept unchanged, part of the drippers is alternately closed to guide the salt away from the root. In order to solve the above problems, in some examples, the field drippers can be divided into two groups of odd and even or strips to generate an alternating plan. When the predicted sodium ion or chloride ion inhibition approaches the critical value while the pH is normal, the alternating strategy is started. In the alternating period, the lateral change of the salt is tracked by the near-surface electrical conductivity imaging or the portable electrical conductivity needle, and if the lateral outward migration reaches the standard, the uniform drip irrigation is restored. The maximum continuous closing time is set to avoid local root over-drying. Thus, the effective salinity of the root zone can be reduced without additional water amount.

[0137] Referring to Figure 11 , an embodiment of the crop irrigation amount prediction device in a drought and saline area in the embodiment of the application can include:

[0138] The calculation unit 21 is configured to calculate the potential evapotranspiration of the target crop based on the predicted key meteorological parameters of the target area by a dual-source evapotranspiration mechanism model, and dynamically update the leaf area index and plant height according to the accumulated temperature process;

[0139] The correction unit 22 is configured to construct different stress inhibition functions for the concerned salt ions and the acid-base degree in the soil solution of the saline-alkali soil, respectively, apply a composite inhibition at the stomatal conductance and / or apparent evapotranspiration resistance site, and correct the potential evapotranspiration to obtain an ideal crop evapotranspiration amount corrected by the salt and the acid-base degree;

[0140] The prediction unit 23 is configured to predict the daily irrigation amount and water distribution schedule suggestion in a future preset period based on the corrected ideal crop evapotranspiration amount.

[0141] As Figure 12 shown, the embodiment of the application further provides an electronic device 300, which includes a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor. When the processor 320 executes the computer program 311, the steps of any method for predicting the crop irrigation amount in a drought and saline area are implemented.

[0142] In the specific implementation process, when the computer program 311 is executed by the processor, any embodiment in the corresponding embodiment can be implemented. Figure 1 ​

Claims

1. A method for predicting the amount of water to be used for irrigation of crops in arid and saline regions, characterized in that, The method comprises the following steps: Based on the predicted key meteorological parameters of the target area, the potential evapotranspiration of the target crop is calculated by a dual-source evapotranspiration mechanism model, and the leaf area index and plant height are dynamically updated according to the accumulated temperature process; For the concerned salt ions and pH in the soil solution of saline-alkali soil, different stress inhibition functions are constructed, and a composite inhibition is applied to the potential evapotranspiration at the stomatal conductance and / or apparent evapotranspiration resistance sites to obtain the ideal crop evapotranspiration modified by salt and pH; Based on the modified ideal crop evapotranspiration, the daily irrigation amount and water distribution schedule in the future preset period are predicted. The method further comprises the following steps: Based on the historical meteorological data of the target area, a multi-member rolling weather forecast is performed by a NeuralGCM weather model, and the historical meteorological data includes an ERA5 reanalysis dataset; The prediction results of the NeuralGCM weather model are input into a Gaussian process regression algorithm coupled with a binary enzyme action optimization algorithm to reforecast key meteorological parameters required by the evapotranspiration model and meteorological variables required for calculation of the leaf area index and plant height model; Three sampling depths are defined, and a stratification factor is established to infer the effective osmotic potential of each layer based on the conductivity and ion concentration; When the salt crust on the surface layer reaches a threshold value and the root layer is relatively low in salt, only the resistance at the soil surface site is inhibited, and the stomatal site is not affected by the salt crust; After the net radiation is allocated according to the leaf area index, the stratification factor is preferentially applied to the soil component to prevent the soil side attenuation from being wrongly attributed to the canopy; The difference between the soil surface temperature rise and the leaf temperature rise in the noon period is used for rapid diagnosis, and if the soil surface temperature rise is significantly higher than the leaf temperature, it is confirmed that the soil site inhibition is dominant, so as to avoid excessive water distribution caused by channel misjudgment.

2. The method of claim 1, wherein, The stress inhibition function comprises at least two of the following: A threshold-slope type inhibition function for sodium ions, a double-threshold interval inhibition function for chloride ions, an optimal-toxic concentration window inhibition function for sulfate ions, and a nonlinear inhibition function for pH.

3. The method of claim 2, wherein, The composite inhibition is applied to the potential evapotranspiration at the stomatal conductance and / or apparent evapotranspiration resistance sites to obtain the ideal crop evapotranspiration modified by salt and pH, which comprises: The stress inhibition functions are combined in parallel or series to form a composite inhibition coefficient at the stomatal conductance and / or apparent evapotranspiration resistance sites to modify the potential evapotranspiration and obtain the ideal crop evapotranspiration modified by salt and pH.

4. The method of claim 1, wherein, The method further comprises the following steps: The environmental stress and ion stress are separated and modeled by using an improved Jarvis model to modify the ideal crop evapotranspiration.

5. The method of claim 4, wherein, The method further comprises the following steps: The modified ideal crop evapotranspiration is evaluated on a daily scale by using Bowen ratio energy balance or eddy correlation flux observation.

6. A device for predicting the amount of water to be used for irrigation of crops in arid and saline regions, characterized in that, The device comprises: A calculation unit is configured to calculate the potential evapotranspiration of the target crop by a dual-source evapotranspiration mechanism model based on the predicted key meteorological parameters of the target area, and dynamically update the leaf area index and plant height according to the accumulated temperature process. The correction unit is configured to construct different stress inhibition functions for the salt ions and the pH value of the saline-alkali soil solution respectively, and to correct the potential evapotranspiration amount by applying a composite inhibition at the stomatal conductance and / or apparent evapotranspiration resistance site, so as to obtain an ideal crop evapotranspiration amount corrected by the salt ions and the pH value; The prediction unit is configured to predict the daily irrigation amount and water distribution schedule within a future preset period based on the ideal crop evapotranspiration amount.

7. An electronic device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor is configured to implement the steps of the crop irrigation amount prediction method in the arid and saline-alkali region according to any one of claims 1-5 when executing the computer program stored in the memory.

8. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executable on the processor to implement the crop irrigation amount prediction method in the arid and saline-alkali region according to any one of claims 1-5.

Images