List of Titles

Investigating the feasibility of using remote sensing in index-based crop insurance for South Africa’s smallholder farming systems

- Masiza, Wonga https://orcid.org/0000-0002-6224-3812

Authors: Masiza, Wonga https://orcid.org/0000-0002-6224-3812
Date: 2021-10
Subjects: Precision farming , Agricultural engineering , Climatic changes
Language: English
Type: Doctoral theses , text
Identifier: http://hdl.handle.net/10353/23000 , vital:54890
Description: Crop farming in Sub-Saharan Africa (SSA) is largely practiced by resource-poor farmers under rain-fed and unpredictable weather conditions. Since agriculture is the mainstay of SSA’s economy, the lack of improved and adapted agricultural technologies in this region sets back economic development and the fight against poverty. Overcoming this constraint and achieving the sustainable development goal to end poverty, requires innovative tools that can be used for weather risk management. One tool that has been gaining momentum recently is index-based crop insurance (IBCI). Since the launch of the first IBCI program in Africa around 2005, the number of IBCI programs has increased. Unfortunately, these programs are constrained by poor product design, basis risk, and low uptake of contracts. When these issues were first pointed-out in the earliest IBCI programs, many reports suggested satellite remote sensing (RS) as a viable solution. Hence, the first objective of this study was to assess how RS has been used in IBCI, the challenges RS faces, and potential contributions of RS that have not yet been meaningfully exploited. The literature shows that IBCI programs are increasingly adopting RS. RS has improved demarcation of unit areas of insurance and enabled IBCI to reach inaccessible areas that do not have sufficient meteorological infrastructure. However, the literature also shows that IBCI is still tainted by basis risk, which emanates from poor contract designs, the influence of non-weather factors on crop yields, imperfect correlations between satellite-based indices and crop yields, and the lack of historical data for calibration. Although IBCI reports cover vegetation and crop health monitoring, few to none cover crop type and crop area mapping. Furthermore, areas including high-resolution mapping, data fusion, microwave RS, machine learning, and computer vision have not been sufficiently tested in IBCI. The second objective of this study was to assess how RS and machine learning techniques can be used to enhance the mapping of smallholder crop farming landscapes. The findings show that machine learning ensembles and the combination of optical and microwave data can map a smallholder farming landscape with a maximum accuracy of 97.71 percent. The third objective was to identify factors that influence crop yields and crop losses in order to improve IBCI design. Results demonstrated that the pervasive notion that low yields in smallholder farms are related to rainfall is an oversimplification. Factors including fertilizer use, seed variety, soil properties, soil moisture, growing degree-days, management, and socioeconomic conditions are some of the most important factors influencing crop yields and crop losses in smallholder farming systems. This shows why IBCI needs to be part of a comprehensive risk management system that understands and approaches smallholder crop farming as complex by linking insurance with advisories and input supplies. Improved inputs and good farming practices could reduce the influence of non-weather factors on crop losses, and thereby reduce basis risk in weather-based index insurance (WII) contracts. The fourth objective of this study was to assess how well the combination of synthetic aperture radar (SAR) and optical indices estimate soil moisture. As stated earlier, soil moisture was found to be one of the most important factors affecting crop yields. Although this method better estimated soil moisture over the first half of the growing season, estimation accuracies were comparable to those found in studies that had used similar datasets (RMSE = 0.043 m3 m-3, MAE = 0.034 m3 m- 3). Further interrogation of interaction effects between the variables used in this study and consideration of other factors that affect SAR backscatter could improve the method. More importantly, incorporating high-resolution satellite-based monitoring of soil moisture into IBCI could potentially reduce basis risk. The fifth objective of this study was to develop an IBCI for smallholder crop farming systems. The proposed IBCI scheme covers maize and derives index thresholds from crop water requirements and satellite-based rainfall estimates. It covers rainfall deficits over the vegetative, mid-season, and late-season stages of maize growth. The key contribution of this system is the derivation of index thresholds from CWR and site-specific rainfall conditions. The widely used approach, which calibrates IBCI by correlating yields and rainfall, exposes contracts to basis risk because, by simply correlating yield and rainfall data, it overlooks the influence of non-weather factors on crop yields and losses. The proposed system must be linked or bundled with non-weather variables that affect crop yields. Effectively, this means that the insurance must be linked or bundled with advisories and input supplies to address the influence of non-weather factors on crop losses. This system also incorporates a crop area-mapping component, which was found to be lacking in many IBCI systems. In conclusion, an IBCI that is based on crop water requirements, which incorporates crop area mapping and links insurance with non-weather crop yield-determining factors, is potentially capable of improving crop insurance for smallholder farming systems. , Thesis (PhD) -- Faculty of Science and Agriculture, 2021
Full Text:
Date Issued: 2021-10

Quick View

Showing items 1 - 2 of 2