AgroMetLLM: An evapotranspiration and agro-advisory system using localized large language models in resource-constrained edge

Authors

  • PARTHA PRATIM RAY Department of Computer Applications, Sikkim University, Gangtok 737102, Sikkim
  • MOHAN PRATAP PRADHAN Department of Computer Applications, Sikkim University, Gangtok 737102, Sikkim

DOI:

https://doi.org/10.54386/jam.v27i3.3081

Keywords:

AgroMetLLM, Evapotranspiration, Raspberry Pi 4B, Quantized LLMs, Ollama Inference, Open-Meteo API

Abstract

We introduce AgroMetLLM, an on-device agrometeorological advisory system that combines five validated evapotranspiration (ET) models with various quantized Large Language Models (LLMs) on a Raspberry Pi 4B. The users specify a location, 3-7-day horizon by using Gradio interface and LLM; Open-Meteo APIs then supply daily inputs Tmax, Tmin, Tmean, RHmean precipitation, Food and Agriculture Organization (FAO) reference evapotranspiration (ET0), and Rs for multiple Indian sites. Computed ET ranges (mm day⁻¹) across locations were: FAO ET₀ 2.84-6.21; Hargreaves-Samani 6.28-13.74; Turc 0.17-0.21; Priestley-Taylor 5.64-9.06; Makkink 2.73-4.38. A few-shot prompting strategy, based on curated examples for 3-, 5-, and 7-day forecasts, is used to guide the Qwen LLM under Ollama to produce structured, five-point advisories in 1-2 s. One-way ANOVA (F = 3.30-6.71, p = 0.016-0.0002) and Kruskal-Wallis tests (χ2 = 9.61-15.48, p < 0.05 except Turc p = 0.088) confirm significant ET differences among models and LLM sizes. All outputs and metadata persist in SQLite, and Matplotlib renders comparative bar charts in the dashboard. These results demonstrate that compact, quantized LLMs can reliably deliver actionable irrigation guidance-matching cloud-based accuracy-while operating offline, with minimal latency and energy use, thus empowering resource-constrained smallholder farmers.

References

Alavi, M., Albaji, M., Golabi, M., Naseri, A.A. and Homayouni, S. (2024). Estimation of sugarcane evapotranspiration from remote sensing and limited meteorological variables using machine learning models. J. Hydrol., 629: 130605.

Al Nazi, Z., Hossain, M.R. and Al Mamun, F. (2025). Evaluation of open and closed-source LLMs for low-resource language with zero-shot, few-shot, and chain-of-thought prompting. Nat. Lang. Process. J., 10: 100124.

Bijlwan, Amit, Shweta Pokhriyal, Rajeev Ranjan, R.K. Singh and Ankita Jha. (2024). Machine learning methods for estimating reference evapotranspiration. J. Agrometeorol., 26(1): 63-68. https://doi.org/10.54386/jam.v26i1.2462

Bakr, D.I., Al-Khalidi, J. and Abed, H.N. (2025). Estimation of climatological parameters using ANN and WEKA models in Diyala Governorate, Iraq. J. Agrometeorol., 27(2): 210-215. https://doi.org/10.54386/jam.v27i2.2877

Jan, S., Khan, U., Khalil, A., Khan, A.A., Jan, H.A. and Ullah, I. (2024). Use of machine learning techniques in predicting inflow in Tarbela reservoir of Upper Indus Basin. J. Agrometeorol., 26(4): 501-504. https://doi.org/10.54386/jam.v26i4.2676

Jiu, T.A., Wei, W.A., Min, X.U., Ning, H.U., Shang, C.H., Xu, J., Hu, X., Lv, H., Z.H. Zi-han and Yu-jing, L.A. (2024). Simulation of evapotranspiration in a well-facilitated paddy field based on machine learning algorithms and energy balance closure. Chin. J. Agrometeorol., 45(10): 1109.

Jhajharia, K. (2025). Wheat yield prediction of Rajasthan using climatic and satellite data and machine learning techniques. J. Agrometeorol., 27(1): 63-66. https://doi.org/10.54386/jam.v27i1.2807

Mayani, M.B. and Itagi, R.L. (2024). Sugarcane acreage estimation using satellite imagery and machine learning. J. Agrometeorol., 26(4): 505-508. https://doi.org/10.54386/jam.v26i4.2669

Naik, R., Majhi, B. and Naidu, D. (2025). Prediction of pan evaporation in Chhattisgarh using machine learning techniques. J. Agrometeorol., 27(1): 86-91. https://doi.org/10.54386/jam.v27i1.2731

Ollama. (2025). Ollama: Run LLMs locally. https://ollama.com/ (Accessed: June 14, 2025)

Open‑Meteo. (2025). Open‑Meteo: Free weather forecasts via API. https://open-meteo.com/ (Accessed: June 14, 2025)

Patel, R.M. and Bunkar, K. (2025). Soybean yield prediction leveraging advanced ensemble machine learning models. J. Agrometeorol., 27(2): 227-229. https://doi.org/10.54386/jam.v27i2.2971

Rajput, J., Singh, M., Lal, K., Khanna, M., Sarangi, A., Mukherjee, J. and Singh, S. (2024). Data-driven reference evapotranspiration (ET₀) estimation: A comparative study of regression and machine learning techniques. Environ. Dev. Sustain., 26(5): 12679-12706.

Shaloo, Kumar, B., Bisht, H., Rajput, J., Mishra, A.K., T.M., K.K. and Brahmanand, P.S. (2024). Reference evapotranspiration prediction using machine learning models: An empirical study from minimal climate data. Agron. J., 116(3): 956-972.

Singh, Y.P., Singh, P.K. and Tomar, A.S. (2024). Performance comparison of linear regression and ANN models in estimating monthly reference evapotranspiration (ET₀). J. Agrometeorol., 26(3): 383-386. https://doi.org/10.54386/jam.v26i3.2641

Downloads

Published

01-09-2025

How to Cite

RAY, P. P., & PRADHAN, M. P. (2025). AgroMetLLM: An evapotranspiration and agro-advisory system using localized large language models in resource-constrained edge. Journal of Agrometeorology, 27(3), 320–326. https://doi.org/10.54386/jam.v27i3.3081

Issue

Vol. 27 No. 3 (2025): September

Section

Research Paper

Categories

License

Copyright (c) 2025 PARTHA PRATIM RAY, MOHAN PRATAP PRADHAN

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

This is a human-readable summary of (and not a substitute for) the license. Disclaimer.

You are free to:

Share — copy and redistribute the material in any medium or format

Adapt — remix, transform, and build upon the material

The licensor cannot revoke these freedoms as long as you follow the license terms.

    Under the following terms:

    Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use.

    NonCommercial — You may not use the material for commercial purposes.

    ShareAlike — If you remix, transform, or build upon the material, you must distribute your contributions under the same license as the original.

    No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.

    Notices:

    You do not have to comply with the license for elements of the material in the public domain or where your use is permitted by an applicable exception or limitation.

    No warranties are given. The license may not give you all of the permissions necessary for your intended use. For example, other rights such as publicity, privacy, or moral rights may limit how you use the material.