Applications of the Google Earth Engine and Phenology-Based Threshold Classification Method for Mapping Forest Cover and Carbon Stock Changes in Siem Reap Province, Cambodia

Pathways from research to sustainable development: Insights from ten research projects in sustainability and resilience

Drawing on collective experience from ten collaborative research projects focused on the Global South, we identify three major challenges that impede the translation of research on sustainability and resilience into better-informed choices by individuals and policy-makers that in turn can support transformation to a sustainable future. The three challenges comprise: (i) converting knowledge produced during research projects into successful knowledge application; (ii) scaling up knowledge in time when research projects are short-term and potential impacts are long-term; and (iii) scaling up knowledge across space, from local research sites to larger-scale or even global impact. Some potential pathways for funding agencies to overcome these challenges include providing targeted prolonged funding for dissemination and outreach, and facilitating collaboration and coordination across different sites, research teams, and partner organizations. By systematically documenting these challenges, we hope to pave the way for further innovations in the research cycle.

Impacts of land use and land cover change on ecosystem service values in the Afroalpine area of Guna Mountain, Northwest Ethiopia

Ecosystem service changes caused by land use and land cover change (LULCC) is an important indictor and early warning of ecological changes. However, few attempts have been made to evaluate the effects of LULCC on ecosystem services in the Afroalpine highlands of Northwestern Ethiopia. Therefore, this study aimed to analyze the impacts of LULCC on ecosystem services values in the afro-alpine area of Guna Mountain, Northwestern Ethiopia. Image classification was carried out using Landsat imageries of 1995, 2008, and 2020 following Random Forest algorithm with Google Earth Engine(GEE) based on filtered sample points. A modified benefit transfer method was used to evaluate ecosystem service value (ESV) changes in response to LULCC. The results revealed that the most notable feature of LULCC in the afro-alpine area of Guna Mountain was the expansion of cropland and built-up areas at the expense of grassland, forest, and shrubland. The overall ESV of the study site was estimated at USD 46.97 × 106 in 1995, USD 36.77 × 106 in 2008, and USD 37.19 × 106 in 2020. The net ESVs of the study site declined by USD 9.78 × 106 between 1995 and 2020. The regulating service values accounted for the greatest share, accounting for over 42% in all periods, followed by provisioning and supporting service values, which accounted for over 29% and 13%, respectively, while cultural services accounted for the smallest amount of the total ESV. The ecosystem service value of food production experienced the highest increase. However, the values of the remaining 16 types of ecosystem services declined with varying degrees of reduction over the study periods. The results of this study is necessary for land-use planners and decision-makers who require site-specific information on impacts of LULCC on ecosystem service.

Assessment of Forest Cover Changes in Vavuniya District, Sri Lanka: Implications for the Establishment of Subnational Forest Reference Emission Level

Assessment of forest cover changes is required to establish the forest reference emission level (FREL) at any scale. Due to civil conflict, such assessments have not yet been undertaken in Sri Lanka, especially in the conflict zone. Here, we assessed the forest cover changes in Vavuniya District, Sri Lanka, from 2001 to 2020, using a combination of the Google Earth Engine (GEE) platform and the phenology-based threshold classification (PBTC) method. Landsat 5 TM data for 2001, 2006, and 2010, and Landsat 8 OLI data for 2016 and 2020 were used to classify forest cover by categories, and their related changes could be assessed by four categories, namely dry monsoon forest, open forest, other lands, and water bodies. With an overall average accuracy of 87% and an average kappa coefficient of 0.83, forest cover was estimated at 57.6% of the total land area in 2020. There was an increase of 0.46% per annum for the entire district between 2001 and 2010, but a drastic loss of 0.60% per year was observed between 2010 and 2020. Specifically, the dry monsoon forest lost 0.30%, but open forest gained 3.62% annually over the same period. Loss and gain of forest cover resulted in carbon emissions and removals of 165,306.6 MgCO2 and 24,064.5 MgCO2 annually, respectively, over the same period. Our findings could be used to set the baseline trend of deforestation, based on which, a subnational forest reference emission level can be established as an emission benchmark, against which comparisons of carbon emissions following the implementation of REDD+ activities can be made, and result-based payment can be claimed under the Paris Agreement.

Two-Stepwise Hierarchical Adaptive Threshold Method for Automatic Rapeseed Mapping over Jiangsu Using Harmonized Landsat/Sentinel-2

Rapeseed distribution mapping is a crucial issue for food and oil security, entertainment, and tourism development. Previous studies have used various remote sensing approaches to map rapeseed. However, the time-consuming and labor-intensive sample data used in these supervised classification methods greatly limit the development of large-scale mapping in rapeseed studies. Regarding threshold methods, some empirical thresholding methods still need sample data to select the optimal threshold value, and their accuracies decrease when a fixed threshold is applied in complex and diverse environments. This study first developed the Normalized Difference Rapeseed Index (NDRI), defined as the difference in green and short-wave infrared bands divided by their sum, to find a suitable feature to distinguish rapeseed from other types of crops. Next, a two-stepwise hierarchical adaptive thresholding (THAT) algorithm requiring no training data was used to automatically extract rapeseed in Xinghua. Finally, two adaptive thresholding methods of the standalone Otsu and Otsu with Canny Edge Detection (OCED) were used to extract rapeseed across Jiangsu province. The results show that (1) NDRI can separate rapeseed from other vegetation well; (2) the OCED-THAT method can accurately map rapeseed in Jiangsu with an overall accuracy (OA) of 0.9559 and a Kappa coefficient of 0.8569, and it performed better than the Otsu-THAT method; (3) the OCED-THAT method had a lower but acceptable accuracy than the Random Forest method (OA = 0.9806 and Kappa = 0.9391). This study indicates that the THAT model is a promising automatic method for mapping rapeseed.

Analysis of Regional Distribution of Tree Species Using Multi-Seasonal Sentinel-1&2 Imagery within Google Earth Engine

Accurate information on tree species is in high demand for forestry management and further investigations on biodiversity and environmental monitoring. Over regional or large areas, distinguishing tree species at high resolutions faces the challenges of a lack of representative features and computational power. A novel methodology was proposed to delineate the explicit spatial distribution of six dominant tree species (Pinus tabulaeformis, Quercus mongolia, Betula spp., Populus spp., Larix spp., and Armeniaca sibirica) and one residual class at 10 m resolution. Their spatial patterns were analyzed over an area covering over 90,000 km2 using the analysis-ready large volume of multisensor imagery within the Google Earth engine (GEE) platform afterwards. Random forest algorithm built into GEE was used together with the 20th and 80th percentiles of multitemporal features extracted from Sentinel-1/2, and topographic features. The composition of tree species in natural forests and plantations at the city and county-level were performed in detail afterwards. The classification achieved a reliable accuracy (77.5% overall accuracy, 0.71 kappa), and the spatial distribution revealed that plantations (Pinus tabulaeformis, Populus spp., Larix spp., and Armeniaca sibirica) outnumber natural forests (Quercus mongolia and Betula spp.) by 6% and were mainly concentrated in the northern and southern regions. Arhorchin had the largest forest area of over 4500 km2, while Hexingten and Aohan ranked first in natural forest and plantation area. Additionally, the class proportion of the number of tree species in Karqin and Ningcheng was more balanced. We suggest focusing more on the suitable areas modeling for tree species using species’ distribution models and environmental factors based on the classification results rather than field survey plots in further studies.