Currently, irrigation decisions in coffee cultivation primarily rely on empirical knowledge, resulting in inefficient practices. Combining real-time leaf water content (LWC) data can help improve the accuracy of the irrigation planning. Spectral remote sensing is a fast, reliable, and non-invasive method to detect vegetation moisture content. In this study, a model to estimate the LWC of Coffea arabica L. was built using hyperspectral reflectance of the canopy under various irrigation levels. For this purpose, common spectral indices, two-band spectral indices [ratio spectral index (RSI); difference spectral index (DSI); and normalized difference spectral index (NDSI)], and three-band spectral indices were constructed. Feature bands were extracted using the successive projections algorithm (SPA). Optimal spectral indices were extracted using the correlation coefficient method, and the feature wavebands and spectral indices were combined into five datasets. These datasets were split into modeling and validation datasets by sample set partitioning based on the joint x-y distance (SPXY) algorithm. A linear model [partial least squares regression (PLSR)] and three non-linear models [support vector machine (SVM); extreme learning machine (ELM); back propagation artificial neural network (BPANN)] were built to estimate LWC of Coffea arabica L. The results indicated that the non-linear models surpassed the linear model. The accuracy was the highest when the modeling was performed using the dataset combination 5. Among various modeling methods, the predictive performance of ELM was the best (modeling dataset: R²=0.745, RMSE=2.241%, RRMSE=3.482%; validation dataset: R²=0.721, RMSE=2.142%, RRMSE=3.364%). ELM outperformed PLSR, SVM, and BPANN in LWC retrieval. The obtained results indicated that the dataset built by the combined use of different methods was superior to those from a single data source in accuracy. This study provides a scientific basis for the quantitative diagnosis of coffee tree water status, with significant implications for optimizing field irrigation management.

Water and fertilizer inputs can dominate the soil quality and crop yield during cultivation in Southwest China. It is essential to explore efficient, collaborative modes in sustainable agriculture. This study aims to systematically evaluated the diverse effects of different irrigation-fertilization modes on the soil water storage, available nutrient content, photosynthesis, edible rose yield, and irrigation water use efficiency (IWUE). A fully factorial design was established, consisting of three irrigation regimes—full irrigation (FI, 100% ETc), light deficit irrigation (DIL, 80% ETc), and moderate deficit irrigation (DIM, 60% ETc)—and four levels of organic fertilizer substitution for chemical fertilizer—R₀ (0%), R₁ (15%), R₂ (30%), and R3 (45%). The treatment FI combined with R₀ served as the control (CK). Correlation analysis and structural equation modeling were applied to explore yield-increasing mechanisms triggered by deficit irrigation with organic fertilizer partial substitution. Experimental results demonstrated that deficit irrigation strategies (DIL and DIM treatments) significantly increased the storage of available soil nutrient reserves by a margin ranging from 6.07% to 15.99%, compared with the full irrigation. Furthermore, the deficit irrigation enhanced the IWUE by 22.76% to 45.00%. The water deficit also caused a measurable decrease in essential physiological indicators. Net photosynthetic rate (Pn), transpiration rate (Tr), stomatal conductance (Gs), and intercellular CO₂ concentration (Ci) were reduced between 10.89% and 40.28%. The DIL treatment achieved an average increase in overall rose yield of 2.23%, with an increase in the total economic benefits of 31.25% during 2023, compared directly to the FI baseline treatment. Conversely, the DIM treatment was reduced the total crop yield and economic benefits over the consecutive two-year period, with the average 7.72% and 24.89%, respectively. Furthermore, the critical parameters—including Pn, Tr, Gs, Ci, harvested crop yield, IWUE, and overall economic benefits—exhibited initially a steady increase, followed by the decrease, as the proportion of organic fertilizer increased progressively. Absolute optimal peak values reached at the R₂ treatment level. Compared with the baseline R₀ treatment, the R₂ treatment shared a remarkable increase in the available reserves of soil nutrient, with the crop yield by 9.64% to 121.56% and 23.72% to 38.52%, respectively, after harvest. Additionally, the IWUE was also enhanced by 32.28% after treatment. The DIMR0 treatment also showed the lowest overall yield, compared with the CK. In contrast, the optimal DILR₂ treatment significantly increased the storage capacity of soil available nutrients, final crop yield, the IWUE, and the highest overall economic benefits. In addition, the structural equation model confirmed that the optimal combination of irrigation and fertilization indirectly regulated the final yield of cultivated edible roses, significantly affecting soil moisture dynamics, nutrient content, and photosynthetic parameters. This finding can provide a highly reliable reference to optimize the irrigation-fertilization systems for edible roses.