In the production of processing spinach (Spinacia oleracea), prespecified fresh weights of produce must be shipped to a processing factory at prespecified times by contract. However, this is not easy, because the growth of spinach, which is raised outdoors, is subject to meteorological conditions. This paper presents SukuSuku Hourensou as a system for predicting and simulating the growth of processing spinach to assist farmers in planning harvests and shipments. SukuSuku Hourensou predicts and simulates spinach growth under meteorological conditions specified by the user after automatically obtaining necessary meteorological data from the Agro-Meteorological Grid Square Data, which was developed and is managed by the National Agriculture and Food Research Organization, Japan. It can adjust prediction and simulation results by using vegetation cover rates obtained from airborne images taken by drones. To support planning of harvests and shipments, SukuSuku Hourensou can suggest harvest times that smooth shipment fresh weights over a harvest season. In the smoothing, SukuSuku Hourensou moves harvest times of farm fields under the condition that the weight of a predicted yield per unit area is included in a range specified by the user. Here, we outline the growth model implemented in SukuSuku Hourensou, describe the procedure for smoothing predicted yields over a harvest season, and present the usage of SukuSuku Hourensou.

Investment in agricultural land and fixed capital goods has become large for reasons of management, scale expansion, adoption of new technologies, improvement of machinery, and other drivers. In evaluating the economic effect and adequacy of investment, it is important to consider not only the profit plan, but also the cash flow, such as raising and repaying funds, and the dynamics of land and fixed capital goods. For a 1-year farming period, a linear programming model is widely used to evaluate the economic efficiency of farm investment. For a 2-year or longer farming period, a dynamic farming plan model is required, such as a multi-stage farming plan model or a recursive farming plan model. Linear programming can handle a multi-stage model, because it seeks an optimal solution for the entire farming period. However, although it struggles to deal with a recursive farming plan model, which seeks an annual optimal solution, a dedicated solution program has not existed. Therefore, we developed a solution program for the recursive programming model for farming activities. The program makes it possible for a user to describe a recursive farming model and to seek the optimal solution of the model in an Excel worksheet.

In the production of high-grade tree fruits, it is important not only to maximize fruit quality, but also to guarantee the quality so that consumers can buy with confidence. It is also necessary to minimize the increase in labor costs associated with maximizing fruit quality. We propose a quality control system named the tutor model for individual high-grade fruits based on information and communication technology. Previously, we have developed a quality control system named the school model for fruits that can be grown in large quantities at a relatively low price, such as mandarin oranges. The school model uses data measured from sample fruits to represent the whole crop. There was no guarantee that the individual fruits meet the required quality standards, and no mechanism that consumers can buy with confidence. In the tutor model, electronic tags or QR code tags are attached to individual fruits. The producer records individual sugar content measured during culture, corresponding to the tags. The system advises the producer to reduce the difference between the target and measured values of individual fruits. Fruits are shipped with their tags so that consumers can obtain information from the tag and buy with confidence. An experiment conducted with ‘Shine Muscat’ grape showed that the risk of shipping low-sugar-content fruits could be reduced and the increase in labor could be restricted to 4.5%.

Extreme weather events pose high risks for agricultural production and cause yield losses in South and Southeast Asia. Tropical cyclones frequently cause significant yield losses in Bangladesh. In this regard, a fuzzy approach for satellite-derived normalized difference vegetation index was used to classify rice yield losses in a coastal region of Bangladesh adjacent to the Bay of Bengal. We used different fuzzy membership functions and overlaid a fuzzification gamma to calculate the expected crisp set to develop the classifiers for yield losses. There were five classes of yield loss: marginal, slight, moderate, very, and extreme. The natural breaks (Jenks) method was used to classify losses as marginal on 461 ha (1.5% of total), 2661 ha (8.6%), moderate 11811 ha (38.2%), very 5814 ha (18.8%), and extreme 10160 ha (32.9%). Field validation identified 29.5% of the reference yield information points in the moderate yield loss class, and 45.2% in extreme. These similarities indicate that the method can be used to estimate yield losses in South and Southeast Asian areas affected by cyclones.