Buddhist temples offer invaluable historical insights into the movement of people across different historical periods and their interactions with the environment. This study focuses on the examination of temple locations in Nakhon Si Thammarat Province, with two primary objectives: analyzing the spatial distribution characteristics of temples across periods to trace the development of temples and settlements, and exploring the relationships between temple locations in each period and physical environmental factors/landforms. To achieve these objectives, GIS tools, including average nearest neighbor (ANN) analysis, kernel density estimation (KDE), and standard deviational ellipse (SDE) analysis, are employed. The study reveals significant patterns. Temples exhibit varying concentrations and shifts over periods, primarily along the ancient trans-peninsular route, with the area around Wat Phra Mahathat Woramahawihan emerging as a prominent center. Over time, temples were built at higher elevations and moving away from waterways. The preference for constructing temples on flat terrains is predominant. In the Sukhothai period (prior to 1349 CE), temples were often situated on flood-resistant sandbars. In the Ayutthaya period (1350 CE – 1767 CE), temples were strategically positioned on plains near waterways and trade routes. During the Thonburi period (1768 CE – 1781 CE), temples were primarily located at higher elevations. From Rattanakosin to present (1782 CE – present), temples expanded to higher elevations due to economic crop cultivation, population growth, and road network development. This study sheds light on the dynamic development of temples and their relationship with environmental factors, contributing to our understanding of historical human settlements and interactions with the landscape.
Field trips play an important role in Geography teaching and learning. During the pandemic, all forms of physical field trips were suspended. While students needed to fulfil curriculum requirements, the switch to virtual field trips was the only option. However, without proper use of technology and pedagogical design for field trips, the teaching and learning experiences of both teachers and students were largely curtailed. This study constructed a virtual platform for field trips in Thailand and Hong Kong. Coupled with important elements of field trip learning, a pedagogical strategy was particularly designed by two geography departments in two universities in Thailand and Hong Kong. The two departments co-organized two field trips, i.e. one virtual and one physical using the same sites in Thailand. To understand the differences in the physical and virtual learning environment from learners’ perspectives, two separate groups of students were asked to evaluate the learning environments from 13 attributes based on the modification of “What Is Happening In this Class? (WIHIC)” Questionnaire. This study focussed on the assessment of the differences between students’ perceptions of preferred and actual learning environment as well as between physical and virtual field trips. The results help (1) identify important learning environment attributes of a meaningful field trip, (2) explore the potential of developing a more effective pedagogy in conducting field trips, and (3) seek synergy between virtual and physical forms of learning environments.
1 January 2026 in Ecosystem Health and Sustainability
Land degradation in palm oil plantations is a critical socioenvironmental challenge in tropical regions. This study develops a Geo-AI (geospatial artificial intelligence) framework integrating GIS and machine learning to generate spatially explicit land degradation susceptibility surfaces for oil palm plantations in Satun Province, Thailand, and to evaluate whether Roundtable on Sustainable Palm Oil (RSPO) certification is associated with lower susceptibility at the plantation scale. The final dataset comprised 10,177 georeferenced observations. Fuzzy logic was used to derive multi-level susceptibility labels from 5 biophysical and management variables, which served as targets for random forest, XGBoost, and support vector machine (SVM) classifiers. Under spatially blocked cross-validation, SVM showed the highest internal consistency [accuracy 97%; receiver operating characteristic (ROC)– area under the curve (AUC) = 0.963]. Although certified plantations represented only 0.9% of observations, nonparametric testing revealed substantially lower susceptibility scores in certified than noncertified plantations (Mann–Whitney U , P < 0.001; Cliff’s δ = −0.96). Global sensitivity and dependency network analyses identified vegetation cover, soil reaction, and slope as the principal interacting determinants of susceptibility. The framework provides interpretable susceptibility intelligence for RSPO auditing, Bio-Circular-Green (BCG)-oriented land-use planning, and targeted risk management. Results should be interpreted in light of class imbalance (60 high-risk observations), a small purposive field sample (9 plots), and context-specific transferability, framing the framework as an extensible susceptibility-assessment tool rather than an externally validated predictive model.