Changes of Wiang Nong Lom and Nong Luang Wetlands in Chiang Saen Valley (Chiang Rai Province, Thailand) During the Period 1988–2017

Pressure on the Wiang Nong Lom and Nong Luang wetland resources in the Chiang Saen Valley of Chiang Rai Province (Thailand) has increased in recent years with the expansion of farmlands and other major sources of wetland conflict related to public land encroachment. Both of these wetlands have been designated as strategic ecosystems. Yet, there is a limited understanding of the way different wetlands respond to change drivers (agriculture, climate, population, etc.), and currently no scientifically valid protocols exist for local wetland mapping and monitoring. Distinguishing between small wetlands and land use and land cover (LULC) components is a challenging affair due to the highly heterogeneous landscape and spectral similarity of compositionally different types of tropical regions. The goals of this article are both technological and substantive, i.e., it aims to (A) propose a synthesis of quantitative techniques that can improve LULC mapping using remotely sensed data (Landsat TM, ETM+, OLI), and (B) assess the wetland changes during the last three decades and better understand the interaction between wetland changes, human population, and the environment. In regards to goal (A), the proposed classification approach employed a synthesis of techniques of decision tree classification (DTC), maximum likelihood classification (MLC), and Mahalanobis distance classification (MDC), with different bands and ancillary data inputs. The results demonstrated that the implementation of the DTC algorithms to address LULC mapping problems exhibited an overall mapping accuracy of 83.9%, which is significantly higher than that of MLC and MDC. It was found that the DTC technique performs well when combined with visible, near-infrared (NIR), and shortwave-infrared bands, a digital elevation model and normalized difference vegetation index layers. Subsequently, the postclassification analysis using DTC showed a notable improvement of approximately 88.0% classification accuracy. Regarding goal (B), our results showed that during the last 30 years, wetland areas in the Chiang Saen Valley have experienced a dramatic decrease of 30.5%, whereas forest landscape surrounding the wetlands has decreased by an astonishing 50.9%. Contrarily, we found that agricultural land size has increased by 24.3%. We suggest that ground data can be linked to the etiology of these results, including the gradual conversion of wetlands to rice cultivation fields as a result of the government rice pledging scheme. Large areas in the study region have been cultivated by para-rubber, palm oil, and pineapple agribusiness production since 2003. In addition, short-term subsidizing government policies promote intensive production for commercial agriculture prompting farmers to transition from subsistence to commercial farming, further impacting wetland conversion. As a result, and in further view of the fact that rapidly expanding agricultural areas have contributed significantly to the decrease of wetland areas during the last three decades, the Chiang Saen Valley wetlands have been designated as wetlands of international importance. The overall recommendation of the present work is that special land-use policy and relevant regulation and/or legislation are critical components of any effort to achieve wetland sustainability.