University of Twente Student Theses


Evaluating the factors influencing farmers’ choices of maize-based cropping patterns and assessing the potential of desis hyperspectral satellite data to discriminate the cropping patterns.

Jepkosgei, Charlynne (2023) Evaluating the factors influencing farmers’ choices of maize-based cropping patterns and assessing the potential of desis hyperspectral satellite data to discriminate the cropping patterns.

[img] PDF
Abstract:Cropping patterns are defined as the annual sequence and spatial arrangement of crops on a piece of land, for example monocropping and intercropping patterns. Information about the location, extent, and types of cropping patterns is crucial for accurately measuring crop production and land use intensity for food security assessment. However, underlying factors that influence farmers' choices about which cropping pattern to adopt are still unclear in most regions in the world. Technology such as remote sensing has enabled the mapping of cropping patterns, though this is challenging in regions dominated by small-holder agriculture in regions like Sub-Saharan Africa due to a variety of limitations such as small field size, highly fragmented landscapes, and the nature of the cropping patterns. This study first explores local knowledge from field farmers’ survey responses by evaluating the factors influencing the choice of monocropped and intercropped maize patterns in Busia, Kenya. It then highlights and discusses six factors, including size of the fields, household needs, availability of resources, farmers’ experience/preference, market demand, pest control/plant symbiosis in comparison to existing literature. Further, the study assesses the discrimination of maize cropping patterns using DESIS hyperspectral satellite data to characterize monocropped and intercropped maize fields. We extracted reflectance of the fields of both cropping patterns based on the field boundary data. Statistical tests identified the bands that showed significant spectral differences between monocropped and intercropped maize fields. A Random Forest (RF) classifier was used for feature selection to identify the best subset of features (bands) that would further be used for classification. The results from the statistical tests indicated a statistical difference in the spectral signatures of the two cropping patterns. As such, 110 significant bands were identified in the visible, red-edge and near-infrared (NIR) spectral regions that were the most sensitive to discriminating the cropping patterns. From feature selection, five bands dominated in the red edge and NIR (752.2nm, 767.5nm, 775.2nm, 783nm, 814.2nm) were further selected for classification. Those bands were used in a RF classifier, obtaining an overall accuracy (OA) of 74% with a producer accuracy (PA) of 71% for monocropped maize fields and 80% for intercropped maize fields and user accuracy (UA) of 91% for monocropped fields and 50% for intercropped fields. An F1 score of 80% for monocropped and 62% for intercropped maize fields was obtained. A kappa coefficient of 0.43 was attained, indicating the complexity of discriminating and classifying the maize-based cropping patterns. The results of this study show that there is potential discrimination of the maize cropping patterns using hyperspectral remote sensing, but it can be quite challenging especially in the late development stage of maize. Hence, remote sensing images need to be obtained during the early part of the growing season before the maize canopy obscures the smaller intercropped crops. The exploratory nature of this research opens more avenues for future research into cropping patterns discrimination in small-holder agriculture and further suggests that a combination of narrative perspectives from farmers and the use of remote sensing technologies is required for an in-depth understanding and addressing current food security challenges. Key words: Farmers’ interview; cropping patterns; hyperspectral data; Mann-Whitney U test; feature selection, random forest.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Subject:38 earth sciences
Programme:Spatial Engineering MSc (60962)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page