Decision Making for Rice Farming Based on Seasonal Forecast System
Selected Entity:JIRCAS
(Japan International Research Center for Agriculture Sciences:JIRCAS)
Project summary
The project is to verify precision of forecast of harvest time of dry season and crop time of rain season crop by seasonal forecast-based decision support system for rice cultivation (WeRise: Weather-rice-nutrient integrated decision support system). At the same time, the project will verify effectivity of advice on farming plans based on WeRise’s forecast information. With proper decision making and farming using WeRise, rice production is expected to increase.
Challenges In irrigated rice cultivation of the target area, costs and waste of water resources due to over-operation of irrigation pumps are major challenges. Accurate forecast of timing and amount of rainfall enables to use stormwater and solve these problems. In the dual cropping, rainfall prediction helps to make an appropriate cultivation plan.
Solution WeRise is an ICT tool that was developed to help rice farmers make decisions. It can make seasonal forecasts up to 12 months ahead and based on that forecast, it can predict the optimum variety of crop, sowing time, fertilization time, and harvest time using a crop growth model. It is applicable for dual cropping and can improve farming plan by supporting to adjust the work schedule between two seasons.
Challenges In irrigated rice cultivation of the target area, costs and waste of water resources due to over-operation of irrigation pumps are major challenges. Accurate forecast of timing and amount of rainfall enables to use stormwater and solve these problems. In the dual cropping, rainfall prediction helps to make an appropriate cultivation plan.
Solution WeRise is an ICT tool that was developed to help rice farmers make decisions. It can make seasonal forecasts up to 12 months ahead and based on that forecast, it can predict the optimum variety of crop, sowing time, fertilization time, and harvest time using a crop growth model. It is applicable for dual cropping and can improve farming plan by supporting to adjust the work schedule between two seasons.

The Status of PoC
Initial Phase
A kick-off meeting was held in February 2021. Weekly meetings are held every Thursday to check progress and exchange information on activity plans. In March, activities on the site began, and based on the results of a questionnaire to farmers, some of the data set necessary for WeRise forecasts was prepared. In July, the shipment of the equipment to collect ground data was completed. In addition, using the existing database of meteorological, soil, and crop data, harvest time and yield prediction using WeRise has started.

Implementation Result
The applicability of the WeRise seasonal forecast in the target area was confirmed through this PoC (Figure 1). In addition, the results indicate that farmers missed the optimal dry season harvest and rainy season sowing times (Figs. 2 and 3), which leads to a decrease in yields and waste of water. These findings confirm the usefulness of WeRise in solving the challenges of improving rice productivity and water resource utilization in Senegal.In the future, it is expected that WeRise will be used in wider areas and contribute to the improvement of rice productivity in Senegal with the increased accuracy through the accumulation of more data and the establishment and strengthening of local human resource development and management systems.

Figure 1. Average error regarding rainfall, average minimum temperature, average maximum temperature, and average wind speed
- It was found that the average error between the weather data observed locally in August 2021 and the weather forecast data of WeRise ( in early June 2021) is relatively small, and the reliability of the forecast information from WeRise can be explained to some extent.

Fig. 2 Relationship between sowing time and harvest time of dry season crops
- Regarding the relationship between the sowing date and the harvest date in the target area, the relationship between the sowing date and the harvest date by the farmers in the target area is classified into the following three groups based on WeRise (130 days of harvest).
☑ Group with shorter growing period and earlier harvest time than WeRise (<WeRise)
☑ Group with similar growth period and harvest time as WeRise (WeRise =)
☑ Group with a longer growing period and a later harvest time than WeRise(WeRise<) - Of the 40 farms, 3 (8%) were <WeRise, 24 (60%) were WeRise <, and WeRise = was 13 (33%).

Fig. 3 Sowing timing and yield prediction for rainy season crops
(Note) SD1: June 27-July 6, SD2: July 7-July 16, SD3: July 17-July 26, SD4: July 27-August 5 , SD5: August 6th-August 15th, SD6: August 16th-August 25th, SD7: August 26th-August 31st
- Figure 3 shows the yield forecast by 2021 rainy season sowing season as of June 2021 using the weather, soil and crop database prepared for the project site.
- WeRise predicts that after SD1-SD2 (June 27-July 17), the yield will begin to decline and will drop below 3t / ha after SD4 (July 27-August 5). It was expected.
- As a result of the interview survey, most of the farmers harvest from mid-July to mid-August, which means that they missed the timing of high yield.
- In this PoC, the sowing time and yield of the rainy season crop were not verified, so it is necessary to verify the predicted value based on the measured values of the rainy season crop.
See announcement of PoC selection for Senegal