Assessing organic and conventional cassava productivity using stochastic frontier production function and crop simulation model

Abstract

Cassava is a major economic crop for millions of small farmers and creates extensive vertical and horizontal value chains. The planted area is spread over the country, especially in the rainfed agroecosystems in the Northeast of Thailand. The majority of small farmers are practicing the conventional cassava (CC) production system while the organic cassava (OC) production system is a viable alternative. However, very few research works using systems approaches have been conducted to understand and predict the transformation into OC by small farmers. Furthermore, little research has been conducted in calibrating and evaluating process-based simulation models and implement models as an agro-informatic tool to assess various agricultural resource management options for transformations under farmers’ situations. This research was undertaken in Yasothon and Khon Kaen provinces, the two major cassava-producing provinces in the Northeast of Thailand during 2018-2022. A combination of household questionnaires, a series of field experiments, and computer simulation model methods were systemically designed and implemented during the research endeavor. The main research objective was to uncover the key components for efficient transformation as well as to provide an agro-informatic tool to assess options for transforming into sustainable cassava production systems. In Yasothon province, the household questionnaires were administered to 496 sampled farmers, with 344 and 152 practicing CC and OC production systems, respectively. The binary logistic regression analysis was applied to explore the promotors and barrier components associated with the transformation. Then, the stochastic frontier production function model was used to evaluate the efficiency and inefficiency factors of cassava productivity for both CC and OC. The results from household questionnaires revealed farmers with CC have been facing with more vulnerability contexts than farmers with OC. Higher income was the main mentioned reason for farmers to make their decision to transform. This reflects the need for structures and process support from public and private institutions to support livelihood strategies at the household level. Policymakers should consider implementing the prospective policies and strategies through relevant institutions on research innovation, extension system, and farmers’ capabilities regarding the sustainable cultural practice transformation. Farmers’ experience, cassava farm size, cassava price, access to extension services, access to credit, and being a member of a farmer group were significantly affected the transformation. Farmers with CC demonstrated obtaining higher fresh storage root yield, income, and profit than farmers with OC. The finding also revealed different efficiency and inefficiency levels for both conventional and organic cassava farming households. An increase in expense on manure and that on tractors decreased the technical efficiency of CC. Off-farm income, crop rotation practice, and the use of the Rayong 7 cassava variety also increased the technical inefficiency of CC. For OC, an increase in expense on manure decreased the technical efficiency of the farming households. Moreover, the use of the Kasetsart 50 variety increased the technical inefficiency of OC. Conventional farmers attained their technical efficiency levels with a mean score of 80 % while organic farmers got a mean technical efficiency score of 90%. These figures suggested that there exists an opportunity for cassava farmers in both groups to improve their technical efficiency; particularly if they can access appropriate technology in terms of manure application, machinery innovation, and crop rotation practice. Besides, the cultivation of site-specific-appropriate cassava variety might enhance both cassava production systems. A series of field experiments was conducted at Khon Kaen Field Crop Research Center, Khon Kaen, Thailand. Two factors consisting of planting dates and cassava varieties were used in the Randomized Complete Block Design (RCBD) experiment with seven replications. The two planting dates were April 2019 and October 2019, while the two cassava varieties were Kasetsart 50 and Rayong 9. A drip irrigation management system was installed to ensure optimal development and growth of cassava during the experimental duration. The minimum data set generated from the field experiment was used to calibrate and evaluate two process-based simulation models. The results yielded a set of calibrated genetic coefficients (GC) for both cassava varieties. The GCs gave a good agreement, based on the agreement index (d-index) and normalized Root Mean Square Error (nRMSE), between the simulated and the observed cassava phenology and growth parameters for CSM-CROPSIM-Cassava and CSM-MANIHOT-Cassava. In addition, the CSM-CROPSIM-Cassava model evaluation provided a good agreement between the simulated and the observed total, leaf, stem, and storage root dry weights with d-index values above 0.8 during the early growth stage. In contrast, the CSM-MANIHOT-Cassava model evaluation overestimated the total crop dry weight of KU50. However, the study suggested the good genetic parameters data sets of the elite cultivars still required additional calibration and evaluation for further applications to assess site-specific technologies for a wide range of climates, soil series, and planting seasons to improve cassava productivity with higher yields and greater economic benefits through recommendations for cassava production system. The model evaluated in this research is an example of an agro-informatic tool that can be used to assess options for transformation into sustainable cassava production systems. As a lesson learned, combining the field survey, field experiments, and crop simulation model methods can provide a unique approach to predict the transformation into OC by small farms and farming performance, expressed as crop productivity, by small farms.