Improving Irrigation Services for Marginal Farmers in India
This post is written by Faisal Hossain, The University of Washington with contributions from Shivam Tripathi, Shahryar Ahmad, Soham Adla, Bharat Lohani, Sandeep Goyal and Karumanchi Sri Harsha
Introduction
In 2016, a collaboration between the University of Washington and Pakistan Council for Research in Water Resources (PCRWR) resulted in the implementation of a cellphone-based irrigation advisory that currently serves 100,000 farmers in the Indus basin (Hossain et al., 2017). The PCRWR Irrigation Advisory is an operational system (see also mirror site) that estimates weekly crop water requirements of different crops at nowcast (last 1 week) and forecast (next 1 week) timescales, using satellite remote sensing techniques to estimate reference evapotranspiration and rainfall of the recent past, and numerical weather prediction forecasts for the future week. This system scaled up from 700 farmers in 2016 to 10,000 in 2017 and now it is serving 100,000 farmers since 2018 in Pakistan. Quantitative impact evaluation of the system by PCRWR revealed that there is about 40 percent savings of irrigation water. With 80 percent of farmers making actionable decisions based on the advisory, this water-saving translates to about 2.5 km3 of groundwater saved each year, which is expected to gradually reverse the trend of decades of groundwater overuse in Pakistan. Building on the successful experience of setting up an irrigation advisory, a pilot project termed Provision of Advisory for Necessary Irrigation (PANI; mirror site) was developed in 2018 for marginal farmers and implemented in 2018 in the Uttar Pradesh state of India.
A Low‐cost Approach to Irrigation Advisory for Marginal Farmers called PANI
The PANI approach builds on the satellite‐based monitoring and forecasting model, which already provides an irrigation advisory service to farmers in Pakistan who own more than 4 hectares of land. Small and marginal farmers own considerably less land (less than 2 hectares), generate barely enough income from selling their crops to survive and are therefore very vulnerable to crop failures. In India, they constitute 86.2 percent of all farmers (Agricultural Census of India, 2015). To target the irrigation advisory to the needs of marginal farmers, additional technology that is low-cost is required. In our pilot for PANI in the Uttar Pradesh, we used sensors to capture field‐specific ground information and used it to downscale coarser, remote sensing‐and weather model‐based advisory that only larger land‐owning farmers can use.
To address the higher spatial resolution needs of marginal farmers at the plot scale, a network of inexpensive, low‐power ground sensors was set up. These sensors can measure air temperature, wind speed and relative humidity at the plot level, as well as soil moisture and soil temperature. The data collected by those sensors is sent to low‐power wide‐area network (LPWAN) gateway towers, which forward the information to a central database (One LPWAN covers a radius of 50 km assuming there are no obstructions between the sensors and the LPWAN tower; de Carvalho Silva, 2017). The central database also downloads freely available satellite and numerical weather prediction model data from the coarser resolution system, which is then fused with ground data to generate plot scale irrigation advisory (Figure 1). Demand for water, or crop water requirements, can be derived knowing the reference evapotranspiration and crop coefficients (a function of crop type and sowing date), whereas supply can be obtained from satellite‐based precipitation estimates. Irrigation requirements are then determined by comparing water demand and water supply.
The PANI system is low‐cost and requires no equipment purchase from the farmers. With a mobile phone that can receive text messages, a farmer can subscribe to basic irrigation information for an estimated US$5 annual fee for the basic version of the service. The system is also low‐maintenance. It operates independently from the electric grid, with the LPWAN towers running on solar power and the ground sensors on batteries.
Figure 1: Schematic showing the main components of PANI based on satellite only irrigation advisory (upper left) that is downscaled using geodatabases (lower right) and cost‐effective Internet of Things (IoT) based ground sensing of soil, weather and crops (upper right).Bringing Marginal Farmers on Board
Extensive engagement with marginal farmers in the pilot area of Uttar Pradesh took place during the launch of the advisory service. The goal of this engagement was to familiarize farmers, understand the context of their irrigation needs, identify demand for new solutions, and work towards engagement and co‐ownership of PANI (Figure 2). Consent and buy‐in from farmers is essential for their support in the deployment and safekeeping of in‐situ sensors, equipment and their overall support in data collection (Figure 3). One‐hundred and fifty wheat (and a few potato) growing farmers were selected for piloting PANI. A baseline survey was carried out to collect data from these farmers regarding farm area, family size, crop types, seed types, fertilizer/pesticide types, irrigation practices, water sources, mobile phone usage, literacy status, land ownership and preferred advisory media (text, voice). Farmers’ plots were digitized for boundaries showcasing many of these survey characteristics.
Figure 2: Interaction with marginal farmers in Uttar Pradesh for PANI piloting in 2018-2019. Figure 3: Deployment of sensors in the fields after consent of farmers, including instruments monitoring soil moisture, electrical conductivity (EC), temperature, and air temperature and humidity.Providing Actionable Information to Marginal Farmers
Climatic parameters used in calculating the reference evapotranspiration (ETo) as a proxy for crop water demand at nowcast timescales were remotely sensed from NASA satellite data. Forecast ETo was derived from the National Ocean and Atmospheric Administration (NOAA)’s Numerical Weather Prediction (NWP) model: the Global Forecast System (GFS; https://www.ncdc.noaa.gov/data‐access/model‐ data/model‐datasets/global‐forcastsystem‐gfs). The necessary computer models and tools were developed for automated computation on a daily basis of the nowcast and forecast of ETo. The ETo was estimated based on the FAO56 report (Allen et al., 1998), which is essentially a modification of the well‐ known Penman‐Monteith equation using temperature, humidity, wind speed and solar radiation as inputs. A web portal was developed (www.i‐pani.com) where the data on ETo as well as additional information such as precipitation (nowcast and forecast) and temperature (forecast) can be downloaded and visualized (Figure 4).
Figure 4. Upper panel – location of the PANI pilot site. Lower panel ‐ Online dashboard for marginal farmer plots participating in the PANI irrigation advisory. The dashboard displays the variables available for visualization for each plot and plot specific information related to soil and crop type, sowing dates, plot size and farmer name.The irrigation advisory presented here is a mobile phone‐based service, which disseminates easily understandable advisories to farmers and collects their feedback. The advisories also communicate weather risks and associated plot scale crop water requirements. Whenever water supply (from precipitation) exceeds crop water demand estimated by the system, farmers are sent messages reassuring that they could use less or no supplemental irrigation. Similarly, when crop water demand is not met by precipitation water supply, farmers are communicated of an irrigation amount they should apply to their plot. The advisories are delivered to farmers’ mobile phones using farmer provided information on sowing date and the type of crop planted (Figure 5)
Figure 5: Illustration of farmer data inputs and samples of weather and irrigation advisory to farmers through SMS.How Well Does PANI Work for Marginal Farmers?
PANI was officially launched during the start of the winter wheat season of October 2018. Irrigation and weather advisory services were provided to farmers via SMS until harvest in March 2019. After harvest, farmers were followed up by our team for a post‐pilot survey. This survey was conducted via phone interviews. Out of the 150 farmers, 69 responded (57.4 percent) and provided valuable feedback on how effective they found PANI to be. Of these, an overwhelming majority indicated that they had read the SMS messages with regularity and about 90 percent were able to understand the messages. Overall, about 85 percent of farmers surveyed indicated either moderate or very large benefit (gains) from using PANI (Figure 6 left panel). This benefit could be in the form of altering irrigation strategy or timing, reducing water, application of fertilizer or using weather information more proactively for land and labor management (Figure 6 right panel).
Conclusion
A first evaluation of the PANI system piloted in Uttar Pradesh showed that the participatory approach to development and implementation of an irrigation advisory was successful in generating buy‐in of marginal farmers, who typically have low confidence in interventions brought from the outside. Our first attempt at piloting PANI for marginal farmers faced with water scarcity revealed the following as take home messages:
- Marginal farmers read the irrigation advisories and weather forecast pushed to their mobile phones daily and found the rainfall and weather forecast to have the most actionable information.
- Marginal farmers reported altering their irrigation timing and amount in response to rainfall forecasts and potato growers, in particular, were eager to minimize wet harvest conditions in the fear of fungal attacks.
- During sowing, temperature forecasts played a more prominent role in farmers’ decision making.
- Significant inertia remains with many farmers for full‐scale adoption of PANI as a service, in particular with marginal farmers, who have a low-risk tolerance as compared with cash crop or fruit growers.
Buoyed by such encouraging feedback from the farmers during the pilot, the PANI system is now poised for a scale‐up in 2020-2021 to serve more marginal farmers of Uttar Pradesh (and potentially other states of India). The timing of our PANI scaling up in India is appropriate, as the Central Government has declared a goal of doubling farmer income for marginal farmers by 2024. Our team has created two summary videos on PANI for dissemination and public engagement. These can be viewed here – PANI-How it works; and PANI farmer feedback. Such videos were found to be very effective in educating potential stakeholders on PANI. In fact, the District Magistrate (DM) of Uttar Pradesh jurisdiction near the pilot site was inspired by our PANI videos and has recently (in December 2019) expressed a desire to have the system expand to 40 more villages and 1000 farmers. The client here is the state government who has also indicated their willingness to pay for the service for each farmer per growing season. We hope to report our experience in scaling up PANI in another report or blog in the near future for our community.
Figure 6. Left panel: Farmer survey during post‐pilot phase on the effectiveness of PANI Right panel: Actions taken by farmers that were triggered by PANI SMS messages.Acknowledgments
The authors wish to acknowledge the contributions in data collection by Anurag Kumar and Neeraj Punetha, instrument deployment by Neeraj Rai and Saurabh, and Yash Gaur in data analysis. Additional support from World Bank Remote Sensing Initiative (Aleix Serrat Capdevilla), University of Washington, Indian Institute of Technology-Kanpur, Kritsnam Technologies and GeoKno are gratefully acknowledged.
References
Agricultural Census of India (2015). All‐India Report on Number and Area of Operational Holdings, Agriculture Census: 2015‐16, Agriculture Census Division, Department of Agriculture, Cooperation & Farmers welfare, Ministry of Agriculture and Farmers welfare, Government of India.
Allen, R. G., Pereira, L. S., Raes, D., & Smith, M. (1998). Crop evapotranspiration‐Guidelines for computing crop water requirements‐FAO Irrigation and drainage paper 56. FAO, Rome, 300(9), D05109.
de Carvalho Silva, J., Rodrigues, J. J., Alberti, A. M., Solic, P., & Aquino, A. L. (2017). LoRaWAN—A low power WAN protocol for Internet of Things: A review and opportunities. In IEEE 2nd International Multidisciplinary Conference on Computer and Energy Science (SpliTech), pp. 1‐6, http://ieeexplore.ieee.org/document/8019271/
Hossain, F., N. Biswas M. Ashraf and Z. Ahmad (2017). Growing more with less using cell phones and satellite data, Eos, vol. 98, https://doi.org/10.1029/2017EO075143.