TY - JOUR

T1 - Distinguishing the windthrow and hydrogeological effects of typhoon impact on agricultural lands

T2 - an integrative OBIA and PPGIS approach

AU - Chen, Tzu-Hsin Karen

AU - Lin, Kuan-Hui Elaine

PY - 2020

Y1 - 2020

N2 - The agricultural impacts of tropical cyclones remain the primary threat to livelihoods in Southeast Asia and Latin America. The impacts take two forms, one is windthrow, i.e . uprooted or snapped trees, caused by strong wind, and the other is hydrogeological effects from heavy precipitation. The empirical effects are different for the two forms. However, little previous research has been devoted to distinguishing the two effects to estimate agricultural losses, and even fewer have used moderate-resolution (10 m resolution) images. This study presents a methodological progression toaddress this deficiency. First, an object-based image analytical method distinguishes the two effects using Satellite Pour l’Observation de la Terre – 5 satellite images. Various objectbased features are compared to acquire their spectral, textural, and geometric characteristics for the interpretation. Second, a public participation geographical information system (PPGIS) approach is developed that combines community empowerment and collaborative field survey to rebuild and represent ground truth during disasters for image validation. The method is applied to a case study of typhoon Bopha that struck Compostela Valley, eastern Mindanao, the Philippines in December 2012. Our assessment indicates that the producer accuracy reaches 88.9% for debris and mud flows and 83.3% for windthrow, and user accuracy reaches 94% and 81%, respectively. The result indicates that the proposed methods have great potential for distinguishing the two effects. It also highlights the efficacy of integrating PPGIS with remote sensing, for image validation purposes and in practice to enhance local residents’environmental consciousness for enhancing adaptive capacity in resource limited regions.

AB - The agricultural impacts of tropical cyclones remain the primary threat to livelihoods in Southeast Asia and Latin America. The impacts take two forms, one is windthrow, i.e . uprooted or snapped trees, caused by strong wind, and the other is hydrogeological effects from heavy precipitation. The empirical effects are different for the two forms. However, little previous research has been devoted to distinguishing the two effects to estimate agricultural losses, and even fewer have used moderate-resolution (10 m resolution) images. This study presents a methodological progression toaddress this deficiency. First, an object-based image analytical method distinguishes the two effects using Satellite Pour l’Observation de la Terre – 5 satellite images. Various objectbased features are compared to acquire their spectral, textural, and geometric characteristics for the interpretation. Second, a public participation geographical information system (PPGIS) approach is developed that combines community empowerment and collaborative field survey to rebuild and represent ground truth during disasters for image validation. The method is applied to a case study of typhoon Bopha that struck Compostela Valley, eastern Mindanao, the Philippines in December 2012. Our assessment indicates that the producer accuracy reaches 88.9% for debris and mud flows and 83.3% for windthrow, and user accuracy reaches 94% and 81%, respectively. The result indicates that the proposed methods have great potential for distinguishing the two effects. It also highlights the efficacy of integrating PPGIS with remote sensing, for image validation purposes and in practice to enhance local residents’environmental consciousness for enhancing adaptive capacity in resource limited regions.

KW - Faculty of Science

KW - Object-based image analysis

KW - PPGIS

KW - Agricultural land use

KW - tropical cyclone

KW - windthrow

KW - Philippines

U2 - 10.1080/01431161.2017.1382741

DO - 10.1080/01431161.2017.1382741

M3 - Journal article

VL - 39

SP - 131

EP - 148

JO - International Journal of Remote Sensing

JF - International Journal of Remote Sensing

SN - 0143-1161

IS - 1

ER -