# Construction, validation, and nitrogen nutrition diagnosis of a critical nitrogen concentration model for cotton

## TL;DR

This study develops and validates a model to diagnose nitrogen needs in cotton plants, helping optimize fertilizer use.

## Contribution

A new critical nitrogen concentration model for cotton using reproductive organ data, offering efficient and accurate nitrogen diagnosis.

## Key findings

- The model based on reproductive organ nitrogen content showed high accuracy with RMSE values of 3.53 and nRMSE of 14.85%.
- Nitrogen nutrition indices (NNI) correlated significantly with seed cotton yield and reached peak performance at 240 kg ha−1 nitrogen application.

## Abstract

Based on continuous field experiments in Jiujiang City, Jiangxi Province, a critical nitrogen concentration dilution curve model for different parts of cotton in the rapeseed cotton rotation pattern in the Yangtze River Basin was constructed. The nitrogen nutrient index (NNI) was used to diagnose the nitrogen abundance and deficiency status, providing a basis for precision fertilization management in cotton fields.

A randomized block design was used, and six nitrogen application treatments (pure nitrogen applications at 0, 60, 120, 180, 240, and 300 kg ha−1) were set up, represented as N0, N60, N120, N180, N240, and N300, respectively. The dry matter mass and nitrogen content of each organ of cotton were measured during the initial flowering stage, peak flowering stage, peak boll stage, and boll opening stage in 2023 and 2024, respectively. A critical nitrogen concentration dilution curve model based on the whole plant nitrogen content and the reproductive organ nitrogen content was constructed using the 2023 data. The model was validated using the 2024 data to compare the nitrogen nutrition diagnostic effects of the two models.

The results showed that the critical nitrogen concentration dilution curve models based on cotton whole plant and reproductive organ biomass (represented by y1 and y2, respectively) were constructed as follows: y1 = 26.936x−0.284 (R2 = 0.6800, P < 0.05), y2 = 23.504x−0.226 (R2 = 0.6226, P < 0.05), the RMSE values were 2.76 and 3.53 respectively, with nRMSE values of 15.10% and 14.85% (both within the range of 10% to 20%), indicating that the model accuracy had reached a good level or above. The nitrogen nutrition index based on the whole plant and reproductive organs (represented by NNI1 and NNI2, respectively) showed a significant positive correlation with seed cotton yield (P < 0.05), and a “linear+plateau” relationship with relative yield (RY) [RY = 1.4879NNI1 − 1.0095 (R2 = 0.6434*), RY = 0.9461NNI2 − 0.3723 (R2 = 0.6165*)]. When the two NNIs reached 1.22 and 1.17, respectively, RY reached its peak. When nitrogen application reached 240 kg ha−1, both NNI values could stably reach 1 or above from the initial flowering stage to the boll opening stage and showed significant advantages in multiple dimensions such as biomass accumulation, yield performance, fiber quality, and nitrogen fertilizer use efficiency.

The critical nitrogen concentration dilution curve model based on the reproductive organ nitrogen content and the nitrogen nutrition diagnosis model had the advantages of being more labor-saving and efficient while maintaining the same level of accurate diagnosis, which could provide practical reference for precision fertilization in cotton fields.