Abstract:

As an important economic crop in Yingshan County, tea is a pillar industry for local rural revitalization. This study aims to determine the localization parameters of the AquaCrop crop model and improve its ability to predict tea yield. Using the tea yields in Yingshan County from 2013 to 2022, as well as meteorological data such as temperature, precipitation, and sunshine, the required tea tree growth parameters for the AquaCrop crop model were obtained through parameter sensitivity analysis. The AquaCrop model parameters were locally calibrated and optimized, and a tea leaf production forecasting model for Yingshan County was established. Based on two future scenario models, SSP2-4.5 and SSP5-8.5, the tea yield was predicted. The results showed that: (1) using OTA method for sensitivity analysis of non-conservative parameters, the relative sensitivity ranking of each parameter was obtained, with maturity time>standardized water productivity>tea tree base temperature>sensitivity of crop coefficient>initial canopy coverage>maximum canopy coverage time, and the sensitivity of other parameters was relatively weak. (2) The localized parameters of the AquaCrop crop model for tea in Yingshan County were optimized, and the root mean square error, conformity, and residual clustering set of the optimized parameter combinations were 0.15, 0.68, and 0.01 respectively. (3) The simulated forecast of tea production in Yingshan County from 2013 to 2022 was basically consistent with the actual production trend. Since 2013, the tea production in Yingshan County had shown an upward trend, reaching 0.03 t/hm². The AquaCrop model predicts an upward trend of 0.02 t/hm², which was basically consistent with the actual production trend. (4) Based on two future scenario models, SSP2-4.5 and SSP5-8.5, the simulated tea yields increased by 13.7% and 38.9% compared to the 2013-2022 average yields due to a significant reduction in tea winter freeze and spring frost and an increase in effective cumulative temperature. In summary, the AquaCrop crop model, after parameter localization, performs well in simulating tea yield in Yingshan County and can serve as a tool for predicting tea yield based on meteorological conditions during the growth period.

Key words: AquaCrop model, OTA, crop model, sensitivity analysis, parameter calibration, tea production forecast, future climate scenario models