Short-medium-long nested water-saving irrigation decision-making method based on numerical weather forecast

Abstract

The invention discloses a short-medium-long nested water-saving irrigation decision-making method based on numerical weather forecast. The method comprises the following steps: (1) checking the precision of a numerical weather forecast product; (2) constructing a real-time irrigation forecasting model; (3) real-time irrigation forecast is driven by numerical weather forecast, and a'short-medium-long 'nested water-saving irrigation decision is implemented; real-time irrigation forecast is driven by numerical weather forecast, a water-saving irrigation decision and an irrigation water plan are implemented layer by layer with the forecast period of weather forecast as a guide, the forecast period of real-time irrigation forecast is prolonged, the problem of rain after irrigation is solved, and the method has a remarkable water-saving effect in the aspect of relieving water resource shortage in the drought period.

Description

Technical Field

[0001] The present invention relates to the technical field of irrigation area water resources planning and management, and in particular to a "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast. Background Art

[0002] As a major agricultural country in the world, my country ranks second in the world in terms of cultivated land area, and agricultural water use accounts for more than 60% of the total water use. At present, the formulation of irrigation water use plans in my country often follows fixed boundary conditions such as regional national economic and social development plans and designed crop planting structures. This inevitably leads to large deviations in water demand forecasts due to dynamic changes in precipitation, resulting in insufficient or excessive irrigation. Therefore, how to utilize limited rainwater resources to minimize agricultural water use while meeting the guarantee rate of crop irrigation is of great research significance for improving the level of water resource utilization in irrigation areas and guiding the allocation of agricultural irrigation water under drought conditions.

[0003] At present, public weather forecasts have been widely used in short- and medium-term crop real-time irrigation forecasts, and their forecast period and accuracy directly affect the effect of irrigation forecasts. Real-time irrigation forecasts determine the irrigation plan and soil moisture dynamics in the future period based on soil moisture conditions, suitable moisture content thresholds, and meteorological data. Compared with traditional conventional decisions, the irrigation risk decision-making method that considers future precipitation can more effectively avoid irrigation waste caused by rain after irrigation, and has been proven to have potential benefits in insufficient irrigation scheduling.

[0004] However, existing studies are mostly based on public weather forecast information for 1-7 days. As the forecast period increases, the accuracy of weather forecasts is often unsatisfactory. If the timeliness of real-time irrigation forecasts is to be improved, the massive weather forecast data with different time steps such as hours, days, and months and different forecast periods such as weeks, months, and seasons in the International Grand Global Ensemble (TIGGE) can be used to further propose a nested irrigation scheme model guided by the forecast period of weather forecasts, which is conducive to improving the continuity and periodicity of the agricultural production process and establishing a more reasonable water-saving irrigation system. Summary of the invention

[0005] Purpose of the invention: The purpose of the present invention is to provide a "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast, which drives real-time irrigation forecast with numerical weather forecast, and implements water-saving irrigation decisions layer by layer and nests irrigation water use plans guided by the forecast period of meteorological forecast, thereby extending the forecast period of real-time irrigation forecast and overcoming the problem of rain after irrigation.

[0006] Technical solution: The "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast described in the present invention comprises the following steps:

[0007] (1) Verify the accuracy of numerical weather forecast products;

[0008] (2) Constructing a real-time irrigation forecast model; including the following steps:

[0009] (21) Estimate crop water requirements;

[0010] (22) Determine the field soil water balance prediction parameters;

[0011] (23) Establish a real-time irrigation forecast model, that is, through the cyclic operation of the field water balance equation, monitor when the soil moisture condition reaches the upper and lower limits of irrigation, so as to determine the irrigation and drainage volume and irrigation and drainage dates;

[0012] (3) Use numerical weather forecasts to drive real-time irrigation forecasts and implement “short-medium-long” nested water-saving irrigation decisions.

[0013] Furthermore, step (1) includes the following steps:

[0014] (11) Using the interactive global ensemble forecast system The International Grand Global Ensemble, the numerical weather forecast products of one or several meteorological forecast centers of the TIGGE data center for the next 1-15 days, using Python to extract grid data and converting point rainfall into area rainfall through the Thiessen polygon method;

[0015] (12) With historical measured precipitation data as a reference, 1 mm / d, 10 mm / d and 20 mm / d were used as the precipitation thresholds for light rain, moderate rain and heavy rain, respectively. The number of correct forecasts, false forecasts, missed forecasts and correct forecasts of no precipitation in the 1-15 day numerical weather forecast products were counted according to the three rainfall levels.

[0016] (13) Four accuracy test indicators, namely, TS score, rain or shine test, underreporting rate, and false alarm rate, are selected to conduct quantitative and qualitative tests on the ensemble forecast products of the forecast center, and the numerical weather forecast products with the best applicability are selected as the input data for real-time irrigation forecasting.

[0017] Furthermore, step (21) is specifically as follows: the reference crop evapotranspiration is estimated according to the FAO56-Penman-Monteith formula, and the crop water requirement is calculated using the single crop coefficient method combined with the linear soil moisture correction coefficient.

[0018] Further, step (22) is specifically as follows: determining the field soil water balance prediction parameters comprises the following steps:

[0019] (221) Obtaining the initial value of the time period: determining the measured soil moisture content or the measured field water layer depth at the beginning of the irrigation forecast period;

[0020] (222) Calculation of effective rainfall: Using the empirical formula method, the effective utilization coefficient of rainfall is α = 0, P∈[0,5]; α = 1, P∈(5,50]; α = 0.9, P∈(50,100]; α = 0.75, P∈(100,150]; α = 0.7, P∈(150,∞).

[0021] (223) Calculation of groundwater recharge: When the groundwater depth is greater than 5 m, it can be considered that there is no groundwater recharge. If the groundwater depth is less than 5 m, the groundwater recharge is calculated using a regression model. The formula is:

[0022] G i =(B-0.15GWD)ET i )

[0023] Where, GWD is the groundwater depth, m; B is the groundwater recharge coefficient;

[0024] (224) Calculate the planned water storage capacity of the wet layer for dry crops: Assuming that the wet soil layer increases linearly with depth, it is related to the measured soil moisture content and is calculated as follows during a certain growth period of the crop:

[0025]

[0026] Where: H n-1 and H n is the planned wet layer depth at the beginning and end of the nth growth stage; t is the number of days in the nth growth stage; γ 水 is the bulk density of water, γ 水 =1g / cm 3 ;

[0027] (225) Determine deterministic forecasting parameters such as crop cultivation system, rice field leakage, field water layer depth control index, suitable moisture content of dryland crops, field water holding capacity, and soil bulk density.

[0028] Furthermore, step (23) is specifically as follows: establish a real-time irrigation forecast model, monitor when the soil moisture condition reaches the upper and lower limits of irrigation through the cyclic operation of the field water balance equation, and determine the irrigation and drainage volume and irrigation and drainage date.

[0029] The field water balance equation for dryland crops is as follows:

[0030] W i+1 =W i +SW i +I i -D i +(Pi +G i -ET i )×0.667×15

[0031] If the soil moisture content of the planned wet layer of dryland crops is lower than the suitable moisture content during the crop growth period, the irrigation calculation is carried out according to the following formula:

[0032]

[0033] If the soil moisture content is higher than the field water holding capacity, drainage is calculated according to the following formula:

[0034]

[0035] Among them, W i , W i+1 are the soil moisture contents at the beginning and end of the time period, m 3 / hm 2 ;P i is the effective rainfall, mm; I i is the irrigation volume, m 3 / hm 2 ;D i is the displacement, m 3 / hm 2 ; G i is the groundwater recharge; ET i is the crop water requirement, mm; SW i is the planned depth of water storage in the wet layer, m 3 / hm 2 θ i is the soil mass moisture content at the end of the period; θ β is the appropriate moisture content; θ tc is the field water holding capacity; γ 土 is the soil dry bulk density, g / cm 3 ; H is the planned wet layer depth, m;

[0036] The field water balance equation for rice is as follows:

[0037] h i+1 =h i +P i +I i -ET i -D i -S i

[0038] If the field water depth drops to the lower limit of the appropriate water depth at the current moment, the irrigation calculation is performed according to the following formula to irrigate to the upper limit of the appropriate water depth;

[0039]

[0040] If the current field water depth rises to more than the maximum allowable water depth, drainage is calculated according to the following formula to drain to the upper limit of the appropriate water depth:

[0041]

[0042] Among them, h i 、h i+1 are the depth of field water layer at the beginning and end of the time period, mm; S i is the leakage of paddy field, mm; h min is the lower limit of suitable water layer depth, mm; h max is the upper limit of suitable water layer depth, mm; h p is the maximum allowable water depth, mm.

[0043] Furthermore, step (3) includes the following steps:

[0044] (31) Inputting the numerical weather forecast product after verification in step (1) into the real-time irrigation forecast model constructed in step (2), performing iterative cycle calculation of the field water balance equation with the weather forecast forecast period as the period, and the calculation period is divided into: short-term 1-3 days, medium-term 1-7 days, and long-term 1-15 days;

[0045] (32) In each calculation cycle, if the rainfall during the forecast period is sufficient to maintain the daily soil moisture content of dryland crops at no less than the appropriate moisture content, there is no need to force irrigation to the field water holding capacity on the forecast day. It is sufficient to ensure that the daily soil moisture content before the arrival of rainfall is no less than the appropriate moisture content; the same applies to rice, ensuring that the daily water layer depth before the arrival of rainfall is no less than the lower limit of the appropriate water layer depth.

[0046] (33) On the forecast day, the short-term irrigation forecast results of 1-3 days, the medium-term irrigation forecast results of 4-7 days, and the long-term irrigation forecast results of 8-15 days are linked together to form a set of "short-medium-long" nested irrigation water use plans for the forecast day. The update cycle of the irrigation water use plan is set to 5 days or 15 days. After each cycle, the soil moisture content and field water layer depth information are updated based on the real-time data of the soil moisture monitoring station, the soil moisture forecast results are supplemented and corrected, the latest 15-day weather forecast data is connected, and the irrigation water use plan is re-drafted based on the real-time irrigation forecast model. This principle is followed until the end of the crop growth period.

[0047] The "short-medium-long" nested water-saving irrigation decision-making system based on numerical weather forecast described in the present invention comprises:

[0048] Verification module: used to verify the accuracy of numerical weather forecast products;

[0049] Irrigation forecast module: used to build a real-time irrigation forecast model; including:

[0050] Estimation module: used to estimate crop water requirements;

[0051] Water balance module: used to determine the field soil water balance prediction parameters;

[0052] Irrigation module: used to establish a real-time irrigation forecast model, that is, to monitor when the soil moisture condition reaches the upper and lower limits of irrigation through the cyclic operation of the field water balance equation, so as to determine the irrigation and drainage volume and irrigation and drainage date;

[0053] Decision-making module: used to drive real-time irrigation forecasts with numerical weather forecasts and implement “short-medium-long” nested water-saving irrigation decisions.

[0054] Furthermore, the inspection module includes the following steps:

[0055] (11) Using the interactive global ensemble forecast system The International Grand Global Ensemble, the numerical weather forecast products of one or several meteorological forecast centers of the TIGGE data center for the next 1-15 days, using Python to extract grid data and converting point rainfall into area rainfall through the Thiessen polygon method;

[0056] (12) With historical measured precipitation data as a reference, 1 mm / d, 10 mm / d and 20 mm / d were used as the precipitation thresholds for light rain, moderate rain and heavy rain, respectively. The number of correct forecasts, false forecasts, missed forecasts and correct forecasts of no precipitation in the 1-15 day numerical weather forecast products were counted according to the three rainfall levels.

[0057] (13) Four accuracy test indicators, namely, TS score, rain or shine test, underreporting rate, and false alarm rate, are selected to conduct quantitative and qualitative tests on the ensemble forecast products of the forecast center, and the numerical weather forecast products with the best applicability are selected as the input data for real-time irrigation forecasting.

[0058] Furthermore, the estimation module is as follows: the reference crop evapotranspiration is estimated according to the FAO56-Penman-Monteith formula, and the crop water requirement is calculated using the single crop coefficient method combined with the linear soil moisture correction coefficient.

[0059] Furthermore, the water balance module is specifically as follows: determining the field soil water balance prediction parameters includes the following steps:

[0060] (221) Obtaining the initial value of the time period: determining the measured soil moisture content or the measured field water layer depth at the beginning of the irrigation forecast period;

[0061] (222) Calculation of effective rainfall: Using the empirical formula method, the effective utilization coefficient of rainfall is α = 0, P∈[0,5]; α = 1, P∈(5,50]; α = 0.9, P∈(50,100]; α = 0.75, P∈(100,150]; α = 0.7, P∈(150,∞).

[0062] (223) Calculation of groundwater recharge: When the groundwater depth is greater than 5 m, it can be considered that there is no groundwater recharge. If the groundwater depth is less than 5 m, the groundwater recharge is calculated using a regression model. The formula is:

[0063] G i =(B-0.15GWD)ET i )

[0064] Where, GWD is the groundwater depth, m; B is the groundwater recharge coefficient;

[0065] (224) Calculate the planned water storage capacity of the wet layer for dry crops: Assuming that the wet soil layer increases linearly with depth, it is related to the measured soil moisture content and is calculated as follows during a certain growth period of the crop:

[0066]

[0067] Where: H n-1 and H n is the planned wet layer depth at the beginning and end of the nth growth stage; t is the number of days in the nth growth stage; γ 水 is the bulk density of water, γ 水 =1g / cm 3 ;

[0068] (225) Determine deterministic forecasting parameters such as crop cultivation system, rice field leakage, field water layer depth control index, suitable moisture content of dryland crops, field water holding capacity, and soil bulk density.

[0069] Furthermore, the irrigation module is specifically as follows: a real-time irrigation forecast model is established to monitor when the soil moisture condition reaches the upper and lower limits of irrigation through the cyclic operation of the field water balance equation, and to determine the irrigation and drainage volume and irrigation and drainage dates.

[0070] The field water balance equation for dryland crops is as follows:

[0071] W i+1 =W i +SW i +I i -D i +(P i +G i -ET i )×0.667×15

[0072] If the soil moisture content of the planned wet layer of dryland crops is lower than the suitable moisture content during the crop growth period, the irrigation calculation is carried out according to the following formula:

[0073]

[0074] If the soil moisture content is higher than the field water holding capacity, drainage is calculated according to the following formula:

[0075]

[0076] Among them, W i , W i+1 are the soil moisture contents at the beginning and end of the time period, m 3 / hm 2 ;P i is the effective rainfall, mm; I i is the irrigation volume, m 3 / hm 2 ;D i is the displacement, m 3 / hm 2 ; G i is the groundwater recharge; ET i is the crop water requirement, mm; SW i is the planned depth of water storage in the wet layer, m 3 / hm 2 θ i is the soil mass moisture content at the end of the period; θ β is the appropriate moisture content; θ tc is the field water holding capacity; γ 土 is the soil dry bulk density, g / cm 3 ; H is the planned wet layer depth, m;

[0077] The field water balance equation for rice is as follows:

[0078] h i+1 =h i +P i +I i -ET i -D i -S i

[0079] If the field water depth drops to the lower limit of the appropriate water depth at the current moment, the irrigation calculation is performed according to the following formula to irrigate to the upper limit of the appropriate water depth;

[0080]

[0081] If the current field water depth rises to more than the maximum allowable water depth, drainage is calculated according to the following formula to drain to the upper limit of the appropriate water depth:

[0082]

[0083] Among them, h i 、h i+1 are the depth of field water layer at the beginning and end of the time period, mm; S i is the leakage of paddy field, mm; h min is the lower limit of suitable water layer depth, mm; h max is the upper limit of suitable water layer depth, mm; h p is the maximum allowable water depth, mm.

[0084] Furthermore, the decision module includes the following steps:

[0085] (31) Inputting the numerical weather forecast product after verification in step (1) into the real-time irrigation forecast model constructed in step (2), performing iterative cycle calculation of the field water balance equation with the weather forecast forecast period as the period, and the calculation period is divided into: short-term 1-3 days, medium-term 1-7 days, and long-term 1-15 days;

[0086] (32) In each calculation cycle, if the rainfall during the forecast period is sufficient to maintain the daily soil moisture content of dryland crops at no less than the appropriate moisture content, there is no need to force irrigation to the field water holding capacity on the forecast day. It is sufficient to ensure that the daily soil moisture content before the arrival of rainfall is no less than the appropriate moisture content; the same applies to rice, ensuring that the daily water layer depth before the arrival of rainfall is no less than the lower limit of the appropriate water layer depth.

[0087] (33) On the forecast day, the short-term irrigation forecast results of 1-3 days, the medium-term irrigation forecast results of 4-7 days, and the long-term irrigation forecast results of 8-15 days are linked together to form a set of "short-medium-long" nested irrigation water use plans for the forecast day. The update cycle of the irrigation water use plan is set to 5 days or 15 days. After each cycle, the soil moisture content and field water layer depth information are updated based on the real-time data of the soil moisture monitoring station, the soil moisture forecast results are supplemented and corrected, the latest 15-day weather forecast data is connected, and the irrigation water use plan is re-drafted based on the real-time irrigation forecast model. This principle is followed until the end of the crop growth period.

[0088] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: (1) It proposes a theoretical framework for real-time irrigation forecast driven by numerical weather forecast, which can effectively extend the forecast period of real-time irrigation forecast, which is conducive to maximizing the use of precipitation resources within the forecast period under the non-full irrigation system and significantly reducing the risk of rain after irrigation. (2) The nested irrigation scheme model guided by the forecast period of meteorological forecast takes into account the continuity and periodicity of the agricultural production process, further improves the efficiency of irrigation water utilization without compromising grain production, and has a significant water-saving effect on guiding the rational allocation of water resources in irrigation areas and alleviating water shortages during drought periods. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 It is a flow chart of the "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast of the present invention;

[0090] Figure 2 It is a schematic diagram of the research area of ​​the present invention;

[0091] Figure 3-1 It is a schematic diagram of the TS scoring test results of the present invention;

[0092] Figure 3-2 It is a schematic diagram of the sunny and rainy test results of the present invention;

[0093] Figure 3-3 It is a schematic diagram of the false negative rate test result of the present invention;

[0094] Figure 3-4 It is a schematic diagram of the empty alarm rate test result of the present invention;

[0095] Figure 4 It is a schematic diagram of the irrigation amount, field water layer depth and precipitation of early rice under different forecast periods of the present invention;

[0096] Figure 5 It is a schematic diagram of the irrigation amount, field water layer depth and precipitation of late rice under different forecast periods of the present invention;

[0097] Figure 6 It is a schematic diagram of the irrigation amount, soil moisture content and precipitation of corn under different forecast periods of the present invention;

[0098] Figure 7 It is a schematic diagram of the irrigation amount, soil moisture content and precipitation of sweet potatoes under different forecast periods of the present invention;

[0099] Figure 8 It is a schematic diagram of irrigation amount, soil moisture content and precipitation of peanuts under different forecast periods of the present invention;

[0100] Fig. 9 It is a schematic diagram of the rolling forecast results of the "short-medium-long" nested water-saving irrigation strategy of the present invention. DETAILED DESCRIPTION

[0101] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0102] like Figure 1 As shown, an embodiment of the present invention provides a "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast, comprising the following steps:

[0103] S10. Accuracy test of numerical weather forecast products;

[0104] S20, constructing a real-time irrigation forecast model;

[0105] S30. Use numerical weather forecasts to drive real-time irrigation forecasts and implement “short-medium-long” nested water-saving irrigation decisions.

[0106] Step S10: numerical weather forecast product accuracy test, specifically including:

[0107] S11. Select the numerical weather forecast products for the next 1-15 days from one or several meteorological forecast centers of the International Grand Global Ensemble (TIGGE) data center, use Python to extract grid data, and convert point rainfall (in grid form) into surface rainfall through the Thiessen polygon method.

[0108] S12. With reference to historical measured precipitation data, 1mm / d, 10mm / d and 20mm / d are used as precipitation thresholds for light rain, moderate rain and heavy rain respectively. The number of correct forecasts, false forecasts, missed forecasts and correct forecasts of no precipitation in 1-15d numerical weather forecast products are counted according to three rainfall levels.

[0109] S13. Select four accuracy test indicators, namely TS score, rain or shine test, missed reporting rate and false reporting rate, to conduct quantitative and qualitative tests on the ensemble forecast products of the forecast center. The calculation formulas for each test indicator are as follows.

[0110] TS=N A / (N A +N B +N C ) (1)

[0111] PC=(N A +N D ) / (N A +N B +N C ) (2)

[0112] PO=N C / (N A+N C ) (3)

[0113] FAR=N B / (N A +N B ) (4)

[0114] Where: TS is the critical success index; PC is the accuracy rate of the weather test forecast; PO is the missed alarm rate; FAR is the false alarm rate. A is the number of correct precipitation forecasts, i.e., rain is forecasted and actually observed; N B N is the number of false reports, i.e., rain is forecast but no rain is actually observed; C is the number of missed reports, i.e., no rain was forecast but rain was actually measured, N D It is the number of correct predictions of no precipitation, that is, no rain is predicted and no rain is actually observed.

[0115] After verification of precipitation ensemble forecast, the numerical weather forecast products with the best applicability are selected as input data for real-time irrigation forecast.

[0116] Step S20 constructs a real-time irrigation forecast model, specifically including:

[0117] S21. Estimate crop water requirements.

[0118] The reference crop evapotranspiration was estimated according to the FAO56-Penman-Monteith formula, and the crop water requirement was calculated using the single crop coefficient method combined with the linear soil moisture correction coefficient. The formula is:

[0119] ET=ET0·K c ·K s (5)

[0120]

[0121] Where: ET is the crop water requirement, mm; ET0 is the reference crop evapotranspiration, mm / d; K s is the soil moisture correction factor; K c is the basic crop coefficient; θ is the soil mass moisture content; θ wp is the wilting moisture content; θ tc is the field water holding capacity.

[0122] Furthermore, the calculation formulas for each variable in the reference crop evapotranspiration are as follows:

[0123]

[0124] R n =R ns -R nl (9)

[0125]

[0126] R s =0.16(T max -T min ) 0.5 R a (12)

[0127] R so =(0.75+2×10 -5 Z)R a (13)

[0128]

[0129]

[0130] Where: Δ is the slope of the saturated water vapor pressure and temperature curve, kPa / ℃; T is the daily average temperature, ℃; R n is the canopy surface net radiation, MJ / (m 2 .d); R ns is the shortwave net radiation; R nl is the long-wave net radiation; R s is the incident radiation on the surface; R so is clear sky radiation; R a is extraterrestrial radiation; d r is the reverse relative distance between the earth and the sun; ω s is the angle at sunset, rad; is latitude, rad; δ is solar declination, rad; Z is altitude, m; n is sunshine hours, h; N is theoretical maximum sunshine hours, h; T max 、T min is the highest and lowest temperature of the day, ℃, and its corresponding Kelvin temperature is T maxk 、T mink , K; J is the day number; a s 、b s is the empirical coefficient; σ is the Stefan-Boltzmann constant, σ=4.903×10 -9 MJ / (K 4 .m 2 .d); G sc is the solar constant, G sc =0.082MJ / (m 2 .min); G is soil heat flux, MJ / (m 2 .d), take 0; γ is the hygrometer constant, kPa / ℃; u2 is the wind speed at 2m above the ground, m / s; e s is the saturated water vapor pressure, kPa; e a is the actual water vapor pressure, kPa, e a=e s ×RH%; RH% is the average relative humidity of air.

[0131] S22. Determine the field soil water balance prediction parameters, including:

[0132] ① Initial value of the period: determined as the actual soil moisture content or the actual field water layer depth at the beginning of the irrigation forecast period.

[0133] ②Effective rainfall: Using the empirical formula method, the effective utilization coefficient of rainfall is α=0, P∈[0,5]; α=1, P∈(5,50]; α=0.9, P∈(50,100]; α=0.75, P∈(100,150]; α=0.7, P∈(150,∞).

[0134] ③ Groundwater recharge: When the local groundwater depth is greater than 5m, it can be considered that there is no groundwater recharge; if the groundwater depth is less than 5m, the regression model can be used to calculate the groundwater recharge volume, the formula is:

[0135] G i =(B-0.15GWD)ET i ) (twenty two)

[0136] Where: GWD is the groundwater depth, m; B is the groundwater recharge coefficient.

[0137] ④ Planned water storage capacity of wet layer for dry crops: Assuming that the depth of wet soil layer increases linearly, it is related to the measured soil moisture content and can be calculated as follows during a certain growth period of the crop:

[0138]

[0139] Where: H n-1 and H n is the planned wet layer depth at the beginning and end of the nth growth stage; t is the number of days in the nth growth stage; γ 水 is the bulk density of water, γ 水 =1g / cm 3 .

[0140] ⑤ Furthermore, it is necessary to determine deterministic forecasting parameters such as crop cultivation system, rice field leakage, field water layer depth control index, suitable moisture content of dry crops, field water holding capacity, and soil bulk density.

[0141] S23. Establish a real-time irrigation forecast model to monitor when the soil moisture condition reaches the upper and lower limits of irrigation through cyclic calculation of the field water balance equation, so as to determine the irrigation and drainage volume and irrigation and drainage dates.

[0142] The field water balance equation of dryland crops is shown in equation (24). If the soil moisture content of the planned wet layer of dryland crops is lower than the suitable moisture content during the crop growth period, the irrigation calculation is performed according to equation (25). If the soil moisture content is higher than the field water holding capacity, the drainage calculation is performed according to equation (26), and the formula is:

[0143] W i+1 =W i +SW i +D i -D i +(P i +G i -ET i )×0.667×15 (24)

[0144]

[0145] Where: W i , W i+1 are the soil moisture contents at the beginning and end of the time period, m 3 / hm 2 ;P i is the effective rainfall, mm; I i is the irrigation volume, m 3 / hm 2 ;D i is the displacement, m 3 / hm 2 ; G i is the groundwater recharge; ET i is the crop water requirement, mm; SW i is the planned depth of water storage in the wet layer, m 3 / hm 2 θ i is the soil mass moisture content at the end of the period; θ β is the appropriate moisture content; θ tc is the field water holding capacity; γ 土 is the soil dry bulk density, g / cm 3 ; H is the planned wet layer depth, m.

[0146] The field water balance equation of rice is shown in equation (27). If the field water layer depth at the current moment drops to the lower limit of the suitable water layer depth, the irrigation calculation is performed according to equation (28) to irrigate to the upper limit of the suitable water depth; if the field water layer depth at the current moment rises to exceed the maximum allowable water depth, the drainage calculation is performed according to equation (29) to drain to the upper limit of the suitable water depth. The formula is:

[0147] h i+1 =h i +P i +I i -ET i -Di -S i (27)

[0148]

[0149] Where: h i 、h i+1 are the depth of field water layer at the beginning and end of the time period, mm; S i is the leakage of paddy field, mm; h min is the lower limit of suitable water layer depth, mm; h max is the upper limit of suitable water layer depth, mm; h p is the maximum allowable water depth, mm.

[0150] S30. Use numerical weather forecasts to drive real-time irrigation forecasts and implement “short-medium-long” nested water-saving irrigation decisions, including:

[0151] S31. Input the numerical weather forecast product after verification in step S10 into the real-time irrigation forecast model constructed in step S20, and perform iterative cycle calculation of the field water balance equation with the weather forecast forecast period as the period. The calculation period is divided into: short-term 1-3 days, medium-term 1-7 days, and long-term 1-15 days.

[0152] S32. In each calculation cycle, if the rainfall in the forecast period is sufficient to maintain the daily soil moisture content of dryland crops at no less than the suitable moisture content, there is no need to force irrigation to the field water holding rate on the forecast day. It is only necessary to ensure that the daily soil moisture content before the rainfall is not less than the suitable moisture content; the same applies to rice, ensuring that the daily water layer depth before the rainfall is not less than the lower limit of the suitable water layer depth.

[0153] S33. On the forecast day, the short-term irrigation forecast results of 1-3 days, the medium-term irrigation forecast results of 4-7 days, and the long-term irrigation forecast results of 8-15 days are linked together to form a set of "short-medium-long" nested irrigation water use plans on the forecast day. The update cycle of the irrigation water use plan is set to 5 days or 15 days. After each cycle, the soil moisture content and field water layer depth information are updated according to the real-time data of the soil moisture monitoring station, the soil moisture forecast results are filled and corrected, the latest 15-day weather forecast data is connected, and the irrigation water use plan is re-drafted based on the real-time irrigation forecast model. According to this principle, it is rolled out until the end of the crop growth period.

[0154] The present invention uses a specific example to illustrate the "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast. This embodiment selects the Huangjiaba Sanhe irrigation area in the Quanmutang irrigation area as an implementation case. The geographical location, meteorological station and soil moisture station of the Huangjiaba Sanhe irrigation area are as follows: Figure 2The Quanmutang Irrigation Area is located in the center of the "Hengshao Drought Corridor" in Hunan Province, occupying a vast hilly area with the Xiangshui and Zishui River watersheds as the central axis. The designed irrigation area is 80,500 hm2. 2 . The irrigation area uses the Quanmutang Reservoir as the main water source, supplemented by more than 300 small and medium-sized reservoirs and more than 18,800 mountain ponds in the irrigation area, forming a long-vine melon-shaped irrigation system that combines storage, diversion and lifting. The irrigation area has a subtropical monsoon climate, with an average annual temperature of 17.43℃ and an average annual precipitation of 1354.56mm. Heavy rains are concentrated in April to August, and seasonal droughts are prone to occur during the peak irrigation seasons in spring and autumn. The Huangjiaba Sanhe Irrigation Area is located in Shaodong City (111°28'E~111°39'E, 27°5'N~27°30'N) in the north-central part of the Quanmutang Irrigation Area. There is one meteorological station (M1) and two soil moisture stations (S1, S2) in the irrigation area, with a designed irrigation area of ​​9,150 hm2. 2 Early rice, late rice, corn, sweet potatoes and peanuts are the main crops in the irrigation area, with an average annual irrigation quota of 4635m 3 / hm 2 .

[0155] Step S10: numerical weather forecast product accuracy test, specifically:

[0156] S11. The numerical weather forecast products of 1-15 days in 2018 from the International Grand Global Ensemble (TIGGE) data center of the China Meteorological Administration (CMA) were selected, and Python was used to extract data and convert point rainfall into surface rainfall through the Thiessen polygon method.

[0157] S12. Taking the measured precipitation data in 2018 as a reference, 1mm / d, 10mm / d and 20mm / d are used as the precipitation thresholds for light rain, moderate rain and heavy rain respectively. The number of correct forecasts, false forecasts, missed forecasts and correct forecasts of no precipitation in the 1-15d numerical weather forecast products are counted according to the three rainfall levels.

[0158] S13, the TS score, rain or shine test, missed report rate, and false report rate are used to test the accuracy of the forecast center's ensemble forecast products. The results show that: TS score value ( Figure 3-1 ) shows a downward trend as the forecast period increases. The forecast effect of the forecast center for light rain is significantly higher than that for other precipitation levels, and the scores within the 7-day forecast period are all over 0.6; the sunny and rainy test ( Figure 3-2 ) shows that the forecast accuracy increases with the increase of rainfall level, among which the forecast effect of moderate rain and heavy rain is the most accurate. Figure 3-3 )、Non-report rate( Figure 3-4) test further analyzed the causes of the errors. Regardless of the level of precipitation, the false alarm rate and the missed alarm rate increased with the extension of the forecast period, and the false alarm rate was higher than the missed alarm rate, which may be related to the overestimation of precipitation in the forecast model. The results of the precipitation ensemble forecast test confirmed that the applicability of CMA numerical weather forecast is better. The forecast results after filling the missing values ​​of false alarms and missed alarms with measured precipitation will be used as the input data for real-time irrigation forecast.

[0159] Step S20 constructs a real-time irrigation forecast model, specifically including:

[0160] S21. Estimate crop water requirements.

[0161] The reference crop evapotranspiration ET0 is estimated according to the FAO56-Penman-Monteith formula, and the crop water requirement is calculated using the single crop coefficient method combined with the linear soil moisture correction coefficient. The statistical results of the crop water requirement for each decade in 2018 are shown in the following table.

[0162]

[0163]

[0164] S22. Determine the field soil water balance prediction parameters, including:

[0165] ① The initial value of the time period is determined as the measured soil moisture content or the measured field water layer depth at the beginning of the irrigation forecast period. In the present invention, since the soil moisture conditions at the beginning of the growth period are unmeasured, the initial value of the time period for dry crops is determined as the appropriate moisture content, and the initial value of the time period for rice is determined as the upper limit of the appropriate water layer depth.

[0166] ②Effective rainfall is calculated using the effective utilization coefficient of precipitation.

[0167] ③ The groundwater in the Huangjiaba Sanhe irrigation area in Example 2 is deeply buried and has serious leakage losses, so it is believed that there is no groundwater recharge.

[0168] ④ The planned depth of the moist layer for dry crops is calculated according to formula (23): The planned moist layer depths at the beginning and end of the growth stage are shown in the following table.

[0169] ⑤ The Huangjiaba Sanhe irrigation area is mainly composed of clay loam and sandy loam, with an average soil bulk density of 1.39 g / cm in the 0-60 cm soil layer. 3 The field water holding capacity (as a percentage of soil mass) is 30.09%.

[0170] ⑥ The deterministic forecast parameters such as crop cultivation system, rice field leakage, field water layer depth control index, suitable moisture content of dry crops, field water holding capacity, crop coefficient, etc. are shown in the following table.

[0171]

[0172] Note: a is the planned wet layer depth, m; b is the suitable moisture content: relative moisture content (as a percentage of soil mass); c is the crop coefficient

[0173]

[0174]

[0175] Note: a is the leakage, mm / d; b is the lower limit of the suitable water layer depth, mm; c is the upper limit of the suitable water layer depth, mm; d is the maximum allowable water depth, mm; e is the crop coefficient.

[0176] S23. Establish a real-time irrigation forecast model, monitor when the soil moisture condition reaches the upper and lower limits of irrigation through the cyclic operation of the field water balance equation, and determine the irrigation and drainage volume and irrigation and drainage dates of dryland crops and rice according to equations (24) to (26) and (27) to (29), respectively.

[0177] S30. Use numerical weather forecasts to drive real-time irrigation forecasts and implement “short-medium-long” nested water-saving irrigation decisions, including:

[0178] S31. Input the numerical weather forecast product after verification in step S10 into the real-time irrigation forecast model constructed in step S20, and perform iterative cycle calculation of the field water balance equation with the weather forecast forecast period as the period. The calculation period is divided into: short-term 1-3 days, medium-term 1-7 days, and long-term 1-15 days.

[0179] S32. According to the water-saving irrigation principle of Example 1, short-term, medium-term and long-term irrigation decisions are made respectively, and irrigation plans for dryland crops and rice in different forecast periods can be obtained.

[0180] The irrigation amount and field water depth changes of early rice and late rice are as follows: Figure 4 , Figure 5 For early rice, the total irrigation volume in the short, medium and long forecast periods is 2857.4m 3 / hm 2 、2506.1m 3 / hm 2 and 2506.1m 3 / hm 2 The average depth of field water layer is 20.31~24.10mm, and the medium- and long-term forecast saves 351.2m compared with the short-term forecast. 3 / hm 2 Among them, although the total amount of irrigation in the medium and long forecast periods is the same, the actual irrigation system is different. For example, during the period from May 21 to May 24, the medium-term forecast adopts a one-time irrigation strategy, and the daily irrigation amount on May 21 reaches 298.6m3 / hm 2 The long-term forecast adopts a strategy of multiple small-scale irrigation, with daily irrigation volume ranging from 63.2 to 98.8 m 3 / hm 2 , but in fact the weather forecast shows that there will be continuous moderate to heavy rain from May 30 to June 5, which means that under the uncertainty of future rainfall, small and frequent irrigation can reduce the risk of rain after irrigation. The short-term forecast is not favorable for heavy rain. For example, the heavy rain on June 20 was as high as 205.5mm, while the short-term forecast on June 15 was 455.4m 3 / hm 2 The irrigation plan has a weak ability to cope with rain after irrigation. For late rice, the total irrigation volume in the short, medium and long forecast periods is 4904.1m 3 / hm 2 、4776.9m 3 / hm 2 、4476.9m 3 / hm 2 The average depth of water layer in the field is 23.6-31.1 mm. The rainfall during the growing period of late rice is significantly less than that during the growing period of early rice. The number of irrigations is more and the amount of irrigation water is also larger.

[0181] The irrigation amount and soil moisture content of corn, sweet potato and peanut are shown as follows: Figure 6 , Figure 7 , Figure 8 In the short, medium and long forecast periods, the irrigation quotas for corn during the 122-day growth period are 2125.8m 3 / hm 2 、1604.5m 3 / hm 2 and 420.5m 3 / hm 2 The predicted irrigation times are 3, 7 and 13 respectively; the irrigation quotas for sweet potatoes during the 214-day growth period are 4557.2m 3 / hm 2 、3567.4m 3 / hm 2 、1752.7m 3 / hm 2 The predicted irrigation times are 11, 29 and 56 respectively; the irrigation quotas for peanuts during the 144-day growth period are 1818.0m 3 / hm 2 、1511.2m 3 / hm 2 、886.5m 3 / hm 2, and the predicted irrigation times are 6, 10, and 33 respectively. Take the irrigation forecast data of corn for 15 consecutive days (June 26 to July 10) as an example. During this period, the measured effective rainfall was 0. On June 30, it was found through the forecast that the soil moisture content was about to fall below the drought threshold (22.56%), and the irrigation program was immediately started. Since the next effective rainfall occurred on July 12, only the long-term forecast could predict the occurrence of this rainfall in advance. Therefore, the short-term, medium-term, and long-term irrigation forecasts were formulated at 586.9m 3 / hm 2 、586.9m 3 / hm 2 、11.5m 3 / hm 2 irrigation plan.

[0182] In general, as the forecast period increases, the method of irrigation with small amounts and multiple times significantly reduces irrigation waste, and the regulatory capacity of irrigation forecasts will become more significant.

[0183] S33. Couple the short-term, medium-term and long-term irrigation forecast results of each crop for 1-3 days, 4-7 days and 8-15 days to form a set of "short-medium-long" nested irrigation water use plans for the forecast day.

[0184] The irrigation areas of early rice, late rice, corn, sweet potato and peanut in the Sanhe irrigation area of ​​Huangjiaba are 4420.7 hm2 respectively. 2 4430.1hm 2 、597.1hm 2 、427.4hm 2 、320.5hm 2 Taking the period from June 29 to July 13, when the development period of dryland crops overlaps with the heading to milk-ripening period of early rice, and the heading to milk-ripening period of late rice from September 16 to September 30 as examples, three water-saving irrigation strategies of 1-15 days under nested forecast periods are given, as shown in the following table.

[0185]

[0186]

[0187] The update cycle of the irrigation water plan is set to 5 days. The soil moisture content and field water depth information are updated based on the real-time data of the soil moisture monitoring station (if any), and the latest 15-day weather forecast data is connected. The irrigation water plan is re-drafted based on the real-time irrigation forecast model until the end of the crop growth period. Taking early rice as an example, the rolling update forecast results of its irrigation water plan are as follows: Fig. 9 shown.

[0188] The embodiment of the present invention provides a "short-medium-long" nested water-saving irrigation decision-making system based on numerical weather forecast, comprising:

[0189] Verification module: used to verify the accuracy of numerical weather forecast products; including the following steps:

[0190] (11) Using the interactive global ensemble forecast system The International Grand Global Ensemble, the numerical weather forecast products of one or several meteorological forecast centers of the TIGGE data center for the next 1-15 days, using Python to extract grid data and converting point rainfall into area rainfall through the Thiessen polygon method;

[0191] (12) With historical measured precipitation data as a reference, 1 mm / d, 10 mm / d and 20 mm / d were used as the precipitation thresholds for light rain, moderate rain and heavy rain, respectively. The number of correct forecasts, false forecasts, missed forecasts and correct forecasts of no precipitation in the 1-15 day numerical weather forecast products were counted according to the three rainfall levels.

[0192] (13) Four accuracy test indicators, namely, TS score, rain or shine test, underreporting rate, and false alarm rate, are selected to conduct quantitative and qualitative tests on the ensemble forecast products of the forecast center, and the numerical weather forecast products with the best applicability are selected as the input data for real-time irrigation forecasting.

[0193] Irrigation forecast module: used to build a real-time irrigation forecast model; including:

[0194] Estimation module: used to estimate crop water requirement; specifically as follows: estimate reference crop evapotranspiration according to the FAO56-Penman-Monteith formula, and calculate crop water requirement using the single crop coefficient method combined with the linear soil moisture correction coefficient.

[0195] Water balance module: used to determine the field soil water balance prediction parameters; specifically as follows: Determining the field soil water balance prediction parameters includes the following steps:

[0196] (221) Obtaining the initial value of the time period: determining the measured soil moisture content or the measured field water layer depth at the beginning of the irrigation forecast period;

[0197] (222) Calculation of effective rainfall: Using the empirical formula method, the effective utilization coefficient of rainfall is α = 0, P∈[0,5]; α = 1, P∈(5,50]; α = 0.9, P∈(50,100]; α = 0.75, P∈(100,150]; α = 0.7, P∈(150,∞).

[0198] (223) Calculation of groundwater recharge: When the groundwater depth is greater than 5 m, it can be considered that there is no groundwater recharge. If the groundwater depth is less than 5 m, the groundwater recharge is calculated using a regression model. The formula is:

[0199] G i =(B-0.15GWD)ET i )

[0200] Where, GWD is the groundwater depth, m; B is the groundwater recharge coefficient;

[0201] (224) Calculate the planned water storage capacity of the wet layer for dry crops: Assuming that the wet soil layer increases linearly with depth, it is related to the measured soil moisture content and is calculated as follows during a certain growth period of the crop:

[0202]

[0203] Where: H n-1 and H n is the planned wet layer depth at the beginning and end of the nth growth stage; t is the number of days in the nth growth stage; γ 水 is the bulk density of water, γ 水 =1g / cm 3 ;

[0204] (225) Determine deterministic forecasting parameters such as crop cultivation system, rice field leakage, field water layer depth control index, suitable moisture content of dryland crops, field water holding capacity, and soil bulk density.

[0205] Irrigation module: used to establish a real-time irrigation forecast model, that is, to monitor when the soil moisture conditions reach the upper and lower limits of irrigation through the cyclic operation of the field water balance equation, so as to determine the irrigation and drainage volume and irrigation and drainage dates; specifically as follows: establish a real-time irrigation forecast model, to monitor when the soil moisture conditions reach the upper and lower limits of irrigation through the cyclic operation of the field water balance equation, so as to determine the irrigation and drainage volume and irrigation and drainage dates.

[0206] The field water balance equation for dryland crops is as follows:

[0207] W i+1 =W i +SW i +D i -D i +(P i +G i -ET i )×0.667×15

[0208] If the soil moisture content of the planned wet layer of dryland crops is lower than the suitable moisture content during the crop growth period, the irrigation calculation is carried out according to the following formula:

[0209]

[0210] If the soil moisture content is higher than the field water holding capacity, drainage is calculated according to the following formula:

[0211]

[0212] Among them, W i , W i+1 are the soil moisture contents at the beginning and end of the time period, m 3 / hm 2 ;P i is the effective rainfall, mm; I i is the irrigation volume, m 3 / hm 2 ;D i is the displacement, m 3 / hm 2 ; G i is the groundwater recharge; ET i is the crop water requirement, mm; SW i is the planned depth of water storage in the wet layer, m 3 / hm 2 θ i is the soil mass moisture content at the end of the period; θ β is the appropriate moisture content; θ tc is the field water holding capacity; γ 土 is the soil dry bulk density, g / cm 3 ; H is the planned wet layer depth, m;

[0213] The field water balance equation for rice is as follows:

[0214] h i+1 =h i +P i +I i -ET i -D i -S i

[0215] If the field water depth drops to the lower limit of the appropriate water depth at the current moment, the irrigation calculation is performed according to the following formula to irrigate to the upper limit of the appropriate water depth;

[0216]

[0217] If the current field water depth rises to more than the maximum allowable water depth, drainage is calculated according to the following formula to drain to the upper limit of the appropriate water depth:

[0218]

[0219] Among them, h i、h i+1 are the depth of field water layer at the beginning and end of the time period, mm; S i is the leakage of paddy field, mm; h min is the lower limit of suitable water layer depth, mm; h max is the upper limit of suitable water layer depth, mm; h p is the maximum allowable water depth, mm.

[0220] Decision-making module: used to drive real-time irrigation forecast with numerical weather forecast and implement "short-medium-long" nested water-saving irrigation decision. It includes the following steps:

[0221] (31) Inputting the numerical weather forecast product after verification in step (1) into the real-time irrigation forecast model constructed in step (2), performing iterative cycle calculation of the field water balance equation with the weather forecast forecast period as the period, and the calculation period is divided into: short-term 1-3 days, medium-term 1-7 days, and long-term 1-15 days;

[0222] (32) In each calculation cycle, if the rainfall during the forecast period is sufficient to maintain the daily soil moisture content of dryland crops at no less than the appropriate moisture content, there is no need to force irrigation to the field water holding capacity on the forecast day. It is sufficient to ensure that the daily soil moisture content before the arrival of rainfall is no less than the appropriate moisture content; the same applies to rice, ensuring that the daily water layer depth before the arrival of rainfall is no less than the lower limit of the appropriate water layer depth.

[0223] (33) On the forecast day, the short-term irrigation forecast results of 1-3 days, the medium-term irrigation forecast results of 4-7 days, and the long-term irrigation forecast results of 8-15 days are linked together to form a set of "short-medium-long" nested irrigation water use plans for the forecast day. The update cycle of the irrigation water use plan is set to 5 days or 15 days. After each cycle, the soil moisture content and field water layer depth information are updated based on the real-time data of the soil moisture monitoring station, the soil moisture forecast results are supplemented and corrected, the latest 15-day weather forecast data is connected, and the irrigation water use plan is re-drafted based on the real-time irrigation forecast model. This principle is followed until the end of the crop growth period.

Claims

1. A "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast, characterized in that: The following steps are involved: (1) Verify the accuracy of numerical weather forecast products; (2) Constructing a real-time irrigation forecast model; including the following steps: (21) Estimate crop water requirements; (22) Determine the field soil water balance prediction parameters; (23) Establish a real-time irrigation forecast model, that is, through the cyclic operation of the field water balance equation, monitor when the soil moisture condition reaches the upper and lower limits of irrigation, so as to determine the irrigation and drainage volume and irrigation and drainage dates; (3) Use numerical weather forecasts to drive real-time irrigation forecasts and implement "short-medium-long" nested water-saving irrigation decisions.

2. The "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast according to claim 1 is characterized in that: Step (1) comprises the following steps: (11) Using the interactive global ensemble forecast system The International Grand Global Ensemble, the numerical weather forecast products of one or several meteorological forecast centers of the TIGGE data center for the next 1-15 days, using Python to extract grid data and converting point rainfall into area rainfall through the Thiessen polygon method; (12) With historical measured precipitation data as a reference, 1 mm / d, 10 mm / d and 20 mm / d were used as the precipitation thresholds for light rain, moderate rain and heavy rain, respectively. The number of correct forecasts, false forecasts, missed forecasts and correct forecasts of no precipitation in the 1-15 day numerical weather forecast products were counted according to the three rainfall levels. (13) Four accuracy test indicators, namely, TS score, rain or shine test, underreporting rate, and false alarm rate, are selected to conduct quantitative and qualitative tests on the ensemble forecast products of the forecast center, and the numerical weather forecast products with the best applicability are selected as the input data for real-time irrigation forecasting.

3. The "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast according to claim 1 is characterized in that: Step (21) is as follows: Estimate the reference crop evapotranspiration according to the FAO56-Penman-Monteith formula, and calculate the crop water requirement using the single crop coefficient method combined with the linear soil moisture correction coefficient.

4. The "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast according to claim 1 is characterized in that: Step (22) is specifically as follows: Determining the field soil water balance prediction parameters includes the following steps: (221) Obtaining the initial value of the time period: determining the measured soil moisture content or the measured field water layer depth at the beginning of the irrigation forecast period; (222) Calculation of effective rainfall: Using the empirical formula method, the effective utilization coefficient of rainfall is α = 0, P∈[0,5]; α = 1, P∈(5,50]; α = 0.9, P∈(50,100]; α = 0.75, P∈(100,150]; α = 0.7, P∈(150,∞). (223) Calculation of groundwater recharge: When the groundwater depth is greater than 5 m, it can be considered that there is no groundwater recharge. If the groundwater depth is less than 5 m, the groundwater recharge is calculated using a regression model. The formula is: G i =(B-0.15GWD)ET i ) Where, GWD is the groundwater depth, m; B is the groundwater recharge coefficient; (224) Calculate the planned water storage capacity of the wet layer for dry crops: Assuming that the depth of the wet soil layer increases linearly, it is related to the measured soil moisture content and is calculated as follows during a certain growth period of the crop: Where: H n-1 and E n is the planned wet layer depth at the beginning and end of the nth growth stage; t is the number of days in the nth growth stage; γ 水 is the bulk density of water, γ 水 =1g / cm 3 ; (225) Determine deterministic forecasting parameters such as crop cultivation system, rice field leakage, field water layer depth control index, suitable moisture content of dryland crops, field water holding capacity, and soil bulk density.

5. The "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast according to claim 1 is characterized in that: Step (23) is as follows: establish a real-time irrigation forecast model, monitor when the soil moisture condition reaches the upper and lower limits of irrigation through the cyclic operation of the field water balance equation, and determine the irrigation and drainage volume and irrigation and drainage date. The field water balance equation for dryland crops is as follows: W i+1 =W i +SW i +I i -D i +(P i +G i -ET i )×0.667×15 If the soil moisture content of the planned wet layer of dryland crops is lower than the suitable moisture content during the crop growth period, the irrigation calculation is carried out according to the following formula: If the soil moisture content is higher than the field water holding capacity, drainage is calculated according to the following formula: Among them, W i , W i+1 are the soil moisture contents at the beginning and end of the time period, m 3 / hm 2 ;P i is the effective rainfall, mm; I i is the irrigation volume, m 3 / hm 2 ;D i is the displacement, m 3 / hm 2 ; G i is the groundwater recharge; ET i is the crop water requirement, mm; SW i is the planned depth of water storage in the wet layer, m 3 / hm 2 θ i is the soil mass moisture content at the end of the period; θ β is the appropriate moisture content; θ tc is the field water holding capacity; γ 土 is the soil dry bulk density, g / cm 3 ; H is the planned wet layer depth, m; The field water balance equation for rice is as follows: h i+1 =h i +P i +1 i -AND i -D i -S i If the field water depth drops to the lower limit of the appropriate water depth at the current moment, the irrigation calculation is performed according to the following formula to irrigate to the upper limit of the appropriate water depth; If the current field water depth rises to more than the maximum allowable water depth, drainage is calculated according to the following formula to drain to the upper limit of the appropriate water depth: Among them, h i 、h i+1 are the depth of field water layer at the beginning and end of the time period, mm; S i is the leakage of paddy field, mm; h min is the lower limit of suitable water layer depth, mm; h max is the upper limit of suitable water layer depth, mm; h p is the maximum allowable water depth, mm.

6. The "short-medium-long" nested water-saving irrigation decision-making method based on numerical weather forecast according to claim 1 is characterized in that: Step (3) comprises the following steps: (31) Inputting the numerical weather forecast product after verification in step (1) into the real-time irrigation forecast model constructed in step (2), performing iterative cycle calculation of the field water balance equation with the weather forecast forecast period as the period, and the calculation period is divided into: short-term 1-3 days, medium-term 1-7 days, and long-term 1-15 days; (32) In each calculation cycle, if the rainfall during the forecast period is sufficient to maintain the daily soil moisture content of dryland crops at no less than the appropriate moisture content, there is no need to force irrigation to the field water holding capacity on the forecast day. It is sufficient to ensure that the daily soil moisture content before the arrival of rainfall is no less than the appropriate moisture content; the same applies to rice, ensuring that the daily water layer depth before the arrival of rainfall is no less than the lower limit of the appropriate water layer depth. (33) On the forecast day, the short-term irrigation forecast results of 1-3 days, the medium-term irrigation forecast results of 4-7 days, and the long-term irrigation forecast results of 8-15 days are linked together to form a set of "short-medium-long" nested irrigation water use plans for the forecast day. The update cycle of the irrigation water use plan is set to 5 days or 15 days. After each cycle, the soil moisture content and field water layer depth information are updated based on the real-time data of the soil moisture monitoring station, the soil moisture forecast results are supplemented and corrected, the latest 15-day weather forecast data is connected, and the irrigation water use plan is re-drafted based on the real-time irrigation forecast model. This principle is followed until the end of the crop growth period.

7. A "short-medium-long" nested water-saving irrigation decision-making system based on numerical weather forecast, characterized in that: include: Verification module: used to verify the accuracy of numerical weather forecast products; Irrigation forecast module: used to build a real-time irrigation forecast model; including: Estimation module: used to estimate crop water requirements; Water balance module: used to determine the field soil water balance prediction parameters; Irrigation module: used to establish a real-time irrigation forecast model, that is, to monitor when the soil moisture condition reaches the upper and lower limits of irrigation through the cyclic operation of the field water balance equation, so as to determine the irrigation and drainage volume and irrigation and drainage date; Decision-making module: used to drive real-time irrigation forecasts with numerical weather forecasts and implement "short-medium-long" nested water-saving irrigation decisions.

8. The "short-medium-long" nested water-saving irrigation decision-making system based on numerical weather forecast according to claim 7 is characterized in that: The estimation module is as follows: the reference crop evapotranspiration is estimated according to the FAO56-Penman-Monteith formula, and the crop water requirement is calculated using the single crop coefficient method combined with the linear soil moisture correction coefficient.

9. The "short-medium-long" nested water-saving irrigation decision-making system based on numerical weather forecast according to claim 7 is characterized in that: The water balance module is as follows: Determine the field soil water balance prediction parameters, including the following steps: (221) Obtaining the initial value of the time period: determining the measured soil moisture content or the measured field water layer depth at the beginning of the irrigation forecast period; (222) Calculation of effective rainfall: Using the empirical formula method, the effective utilization coefficient of rainfall is α = 0, P∈[0,5]; α = 1, P∈(5,50]; α = 0.9, P∈(50,100]; α = 0.75, P∈(100,150]; α = 0.7, P∈(150,∞). (223) Calculation of groundwater recharge: When the groundwater depth is greater than 5 m, it can be considered that there is no groundwater recharge. If the groundwater depth is less than 5 m, the groundwater recharge is calculated using a regression model. The formula is: G i =(B-0.15GWD)ET i ) Where, GWD is the groundwater depth, m; B is the groundwater recharge coefficient; (224) Calculate the planned water storage capacity of the wet layer for dry crops: Assuming that the wet soil layer increases linearly with depth, it is related to the measured soil moisture content and is calculated as follows during a certain growth period of the crop: Where: H n-1 and H n is the planned wet layer depth at the beginning and end of the nth growth stage; t is the number of days in the nth growth stage; γ 水 is the bulk density of water, γ 水 =1g / cm 3 ; (225) Determine deterministic forecasting parameters such as crop cultivation system, rice field leakage, field water layer depth control index, suitable moisture content of dryland crops, field water holding capacity, and soil bulk density.

10. The "short-medium-long" nested water-saving irrigation decision-making system based on numerical weather forecast according to claim 7, characterized in that: The irrigation module is as follows: establish a real-time irrigation forecast model, monitor when the soil moisture condition reaches the upper and lower limits of irrigation through the cyclic operation of the field water balance equation, and determine the irrigation and drainage volume and irrigation and drainage dates. The field water balance equation for dryland crops is as follows: W i+1 =W i +SW i +I i -D i +(P i +G i -ET i )×0.667×15 If the soil moisture content of the planned wet layer of dryland crops is lower than the suitable moisture content during the crop growth period, the irrigation calculation is carried out according to the following formula: If the soil moisture content is higher than the field water holding capacity, drainage is calculated according to the following formula: Among them, W i , W i+1 are the soil moisture contents at the beginning and end of the time period, m 3 / hm 2 ;P i is the effective rainfall, mm; I i is the irrigation volume, m 3 / hm 2 ;D i is the displacement, m 3 / hm 2 ; G i is the groundwater recharge; ET i is the crop water requirement, mm; SW i is the planned depth of water storage in the wet layer, m 3 / hm 2 θ i is the soil mass moisture content at the end of the period; θ β is the appropriate moisture content; θ tc is the field water holding capacity; γ 土 is the soil dry bulk density, g / cm 3 ; H is the planned wet layer depth, m; The field water balance equation for rice is as follows: h i+1 =h i +P i +1 i -AND i -D i -S i If the field water depth drops to the lower limit of the appropriate water depth at the current moment, the irrigation calculation is performed according to the following formula to irrigate to the upper limit of the appropriate water depth; If the current field water depth rises to more than the maximum allowable water depth, drainage is calculated according to the following formula to drain to the upper limit of the appropriate water depth: Among them, h i 、h i+1 are the depth of field water layer at the beginning and end of the time period, mm; S i is the leakage of paddy field, mm; h min is the lower limit of suitable water layer depth, mm; h max is the upper limit of suitable water layer depth, mm; h p is the maximum allowable water depth, mm.

Images