CN-122017198-A - Delta-based15N medlar soil nitrogen circulation system evaluation method

CN122017198ACN 122017198 ACN122017198 ACN 122017198ACN-122017198-A

Abstract

The invention provides a method for evaluating a medlar soil nitrogen circulation system based on delta 15 N, and belongs to the technical field of intelligent agriculture and precise fertilization. The method aims at solving the problems that the traditional medlar fertilization depends on experience and static soil test, and accurate fertilization decision based on a nitrogen circulation mechanism is difficult to realize. The method comprises the steps of setting tests of different nitrogen source treatments, synchronously measuring the ammonium nitrogen content and nitrate nitrogen content of soil, delta 15 N values of the ammonium nitrogen content and the nitrate nitrogen content, soil culture transformation data and delta 15 N values of medlar leaves, obtaining fruit quality indexes, establishing a soil-plant delta 15 N response relation based on the data, calculating nitrogen transformation rate, integrating parameters, constructing a comprehensive evaluation model, outputting soil nitrogen plant effectiveness scores, nitrogen cycle strength scores and fruit quality prediction intervals according to the input soil delta 15 N values, and generating a fertilization decision scheme containing specific nitrogen source types and application amounts based on the evaluation results. The method is mainly used for nitrogen nutrition diagnosis and accurate fertilization management in medlar planting.

Inventors

NIU SHUHUI
JU YANJUN
ZHAO DUOYONG
LIU HEJIANG
KANG LU

Assignees

新疆维吾尔自治区农业科学院

Dates

Publication Date: 20260512
Application Date: 20260228

Claims (9)

1. A method for evaluating a medlar soil nitrogen circulation system based on delta 15 N is characterized by comprising the following steps: In the main growth stage of the medlar, synchronously measuring the ammonium nitrogen and nitrate nitrogen contents of soil under each treatment and delta 15 N values thereof, inorganic nitrogen content change values before and after soil culture for reflecting nitrogen conversion dynamics and delta 15 N values of functional leaves of medlar plants, and obtaining medlar fruit quality indexes of corresponding test cells, wherein the fruit quality indexes comprise soluble solid content, total sugar content, protein content and amino acid content; based on the measured soil inorganic nitrogen delta 15 N value and the medlar leaf delta 15 N value, establishing a quantitative function relation between the soil inorganic nitrogen delta 15 N mean value and the medlar leaf delta 15 N value through regression analysis; calculating the apparent net mineralization rate and apparent net nitrification rate of the soil nitrogen based on the inorganic nitrogen content change values before and after the soil culture; Constructing a comprehensive evaluation model which takes the delta 15 N values of soil ammonium nitrogen and nitrate nitrogen as input and takes the soil nitrogen plant effectiveness quantization score, the nitrogen cycle transformation intensity quantization score and a fruit quality prediction interval as output, wherein the fruit quality prediction interval comprises a prediction range of soluble solid content, total sugar content, protein content and amino acid content; The method comprises the steps of carrying out evaluation on target medlar field blocks based on the comprehensive evaluation model, wherein the evaluation comprises the steps of detecting delta 15 N values of current ammonium nitrogen and nitrate nitrogen of field blocks and inputting the delta 15 N values into the model, so that a soil nitrogen plant effectiveness quantization score, a nitrogen cycle conversion intensity quantization score and a fruit quality prediction interval aiming at the field blocks are obtained, and according to a comparison result of the fruit quality prediction interval and a preset quality target and numerical values of the soil nitrogen plant effectiveness quantization score and the nitrogen cycle conversion intensity quantization score, a fertilization decision scheme comprising suggestions of nitrogen source types and application amount is generated, and the preset quality target comprises reference values of soluble solid content, total sugar content, protein content and amino acid content.
2. The method of claim 1, wherein at least one of the following environmental factor data is synchronously monitored and acquired during the synchronization of the field trials and corresponding periods of evaluation of the target field pieces: Weather data, including average air temperature and rainfall; soil physical data, including soil volume moisture content; And when the comprehensive evaluation model is constructed, the acquired environmental factor data and delta 15 N values of the soil ammonium nitrogen and the nitrate nitrogen are combined to form an input feature set of the model.
3. The method of claim 2, wherein the comprehensive evaluation model is constructed by a gradient-lifting decision tree algorithm; The input characteristics of the gradient lifting decision tree model comprise the soil ammonium nitrogen delta 15 N value, the nitrate nitrogen delta 15 N value, regression coefficients in the quantitative function relationship, the numerical values of the apparent net mineralization rate and the net nitrifying rate and the key environmental factor data; the model is configured to be capable of simultaneously predicting an output soil nitrogen plant effectiveness quantification score, a nitrogen cycle conversion intensity quantification score, and a fruit quality prediction interval based on the input features.
4. A method according to claim 3, wherein the gradient-lifting decision tree model is trained by minimizing a multitasking loss function; The multitasking loss function is a linear weighted sum of prediction error items corresponding to the soil nitrogen plant effectiveness quantization score, the nitrogen cycle transformation intensity quantization score and the fruit quality prediction interval respectively; the weight coefficients in the linear weighted sum are determined by iteratively optimizing the weight coefficients using validation set performance feedback during model training.
5. The method of claim 4, wherein the iteratively optimizing the weight coefficients using validation set performance feedback is determined by performing a process comprising: assigning a group of preset initial values for the weight coefficients, and dividing the field test data set into a training subset and a verification subset which are not overlapped with each other; Performing optimization cycles for a plurality of times, training the gradient lifting decision tree model on the training subset by using a current weight coefficient in each cycle, predicting on the verification subset by using the trained model, respectively calculating the soil nitrogen plant effectiveness quantization score, the nitrogen cycle transformation intensity quantization score and the prediction error of a fruit quality prediction interval, and calculating according to a preset weight updating algorithm and the prediction error to obtain a group of new weight coefficient candidate values; Substituting the new weight coefficient candidate values into a model and re-evaluating comprehensive prediction performance on the verification subset, and updating the weight coefficients into the candidate values if the comprehensive performance is superior to the performance using the current weight coefficients; And repeating the optimization circulation and the evaluation updating process until a preset optimization termination condition is reached, and taking the finally obtained weight coefficient as a determined value in the linear weighted sum.
6. The method according to any one of claims 1 to 5, wherein generating a fertilization decision scheme comprising advice of nitrogen source type and amount applied is accomplished by querying a pre-set decision map knowledge base; the decision mapping knowledge base takes the soil nitrogen plant effectiveness quantitative score, the nitrogen cycle transformation intensity quantitative score and the comparison result of the fruit quality prediction interval and a preset quality target as joint input conditions; The knowledge base is pre-stored with a mapping relation between a plurality of groups of input condition combinations and corresponding output decisions, wherein each group of output decisions comprises specific recommended nitrogen source types and application amount values.
7. The method of claim 6, wherein generating a fertilization decision scheme comprising a nitrogen source type and a fertilization amount recommendation is performed by a trained fertilization scheme recommendation model; The fertilization scheme recommendation model takes the soil nitrogen plant effectiveness quantitative score, the nitrogen cycle transformation intensity quantitative score and the comparison result of the fruit quality prediction interval and a preset quality target as input characteristics; The fertilization scheme recommendation model takes fertilization scheme data which corresponds to the input characteristics and is verified to be preferred in a historical agronomic test as a training label; The output of the model is the recommended nitrogen source type and the application amount value matched with the system state characterized by the input characteristics.
8. The method of claim 7, wherein the fertilization scheme recommendation model is a multi-layer perceptron neural network; the multi-layer perceptron neural network comprises an input layer, at least one hidden layer and an output layer, wherein the output layer is configured to simultaneously output discrete classification results representing different nitrogen source types and continuous values representing recommended application amounts.
9. The method according to claim 5, wherein in the iterative optimization of the weight coefficients using the verification set performance feedback, the preset weight update algorithm is specifically: Dynamically adjusting the weight coefficient of each task in the loss function according to the latest round of prediction error of each task on the verification subset; The method comprises the steps of setting an error reference threshold, increasing the loss weight of a task with a prediction error exceeding the error reference threshold according to an exceeding proportion in the next iteration, wherein the increasing amplitude is in direct proportion to the exceeding proportion, and keeping the loss weight unchanged for the task with the prediction error being lower than the error reference threshold.

Description

Evaluation method of medlar soil nitrogen circulation system based on delta 15 N Technical Field The invention belongs to the technical field of intelligent agriculture and precise fertilization, and particularly relates to a method for evaluating a medlar soil nitrogen circulation system based on delta 15 N. Background In the planting production of medlar, scientific fertilization is a key link for guaranteeing the yield, improving the quality and maintaining the soil health. Traditional medlar fertilization management mainly depends on experience judgment of growers and conventional test of soil basic nutrient content. This model is generally associated with insufficient attention to the dynamic conversion process of soil nitrogen, and especially lacks a real-time and accurate quantitative assessment means for migration, conversion and absorption of nitrogen in a soil-plant system. In order to optimize fertilization, agricultural research and practice have also attempted to introduce finer soil testing and plant nutrition diagnostic techniques. However, conventional soil nitrogen morphological chemical analysis can only provide static content information at a certain point in time, and cannot reveal the conversion rate of nitrogen, source contribution and dynamic process of plant absorption. The judgment of the nitrogen supply capacity of the soil still remains in the appearance level, the differential effect generated by different fertilization measures is difficult to explain mechanically, and the long-term effect after fertilization cannot be predicted. In addition, the nitrogen nutrition status of the medlar plants is traditionally and indirectly estimated mainly through leaf nitrogen content or chlorophyll measurement, but the indexes are easy to be interfered by environmental factors, and the nitrogen is not effectively differentiated from the current-season fertilizer or the original soil warehouse, so that the timeliness and pertinence for guiding accurate topdressing are limited. Disclosure of Invention It is an object of the present invention to address at least the above problems and/or disadvantages and to provide at least the advantages described below. The invention aims to solve the problems that the traditional medlar fertilization depends on experience and static soil test, the dynamic conversion process of soil nitrogen, the plant absorption effectiveness cannot be systematically quantized, and the indexes of the process cannot be correlated with the quality of the final fruits, so that the accurate fertilization decision based on a nitrogen circulation mechanism is difficult to realize. The invention aims to provide a method for evaluating a medlar soil nitrogen circulation system based on delta 15 N, which comprises the following steps: In the main growth stage of the medlar, synchronously measuring the ammonium nitrogen and nitrate nitrogen contents of soil under each treatment and delta 15 N values thereof, inorganic nitrogen content change values before and after soil culture for reflecting nitrogen conversion dynamics and delta 15 N values of functional leaves of medlar plants, and obtaining medlar fruit quality indexes of corresponding test cells, wherein the fruit quality indexes comprise soluble solid content, total sugar content, protein content and amino acid content; based on the measured soil inorganic nitrogen delta 15 N value and the medlar leaf delta 15 N value, establishing a quantitative function relation between the soil inorganic nitrogen delta 15 N mean value and the medlar leaf delta 15 N value through regression analysis; calculating the apparent net mineralization rate and apparent net nitrification rate of the soil nitrogen based on the inorganic nitrogen content change values before and after the soil culture; Constructing a comprehensive evaluation model which takes the delta 15 N values of soil ammonium nitrogen and nitrate nitrogen as input and takes the soil nitrogen plant effectiveness quantization score, the nitrogen cycle transformation intensity quantization score and a fruit quality prediction interval as output, wherein the fruit quality prediction interval comprises a prediction range of soluble solid content, total sugar content, protein content and amino acid content; The method comprises the steps of carrying out evaluation on target medlar field blocks based on the comprehensive evaluation model, wherein the evaluation comprises the steps of detecting delta 15 N values of current ammonium nitrogen and nitrate nitrogen of field blocks and inputting the delta 15 N values into the model, so that a soil nitrogen plant effectiveness quantization score, a nitrogen cycle conversion intensity quantization score and a fruit quality prediction interval aiming at the field blocks are obtained, and according to a comparison result of the fruit quality prediction interval and a preset quality target and numerical values of the soil nitrogen plant effectiveness