Climate change is one of the greatest challenges in Kyrgyzstan. There have been negative spillover effects in agriculture. This study aims to assess the climate change impacts on cropland suitability in Kyrgyzstan. We used the random forest algorithm to develop a model that captures the effects of multiple climate and environment factors at a spatial resolution of 1 km². The model was then applied in the scenario analysis for an understanding of how climate change affects cropland distribution. The potential high-quality cropland was found to be included in existing croplands, while the remaining were distributed around the Chu-Talas valley, the Issyk-kul area, and the Fergana valley. These potential high-quality croplands comprise grasslands (47.1%) and croplands (43.7%). In the future, the potential high-quality cropland exhibited inland trends at the periphery of original cropland category, with grassland and cropland as the primary land components. Due to climate change, potential high-quality cropland is expected to gradually reduce from the 2050s to the 2070s, exhibiting the largest reduction in potential high-quality areas for the Representative Concentration Pathway 8.5 scenario. Therefore, the short- and long-term adaptation strategies are needed for prioritizing the croplands to ensure food security and agricultural resilience.

Land Use, Land-Use Change and Forestry includes the only carbon sink among greenhouse gas (GHG) inventory, and the significance of accurate statistics has increased. However, the exact amounts of GHG absorption and emissions cannot be calculated due to the absence of definition and activity data in the settlement sector. Therefore, this study established the concept and activity data of settlement area in South Korea by reviewing the definition suggested by IPCC and Annex 1 countries and calculated GHG absorption. Considering the possibility of connecting current MRV guidelines, exisiting national statistics, country and urban planning data, and unifying management authority, settlement can be defined in three ways, as a Land category, Urban area, and Urban & Plan management. The study built activity data based on concept 1 and 2, which rarely overlap with the other land categories, and the study area is Seoul, with the largest proportion of settlements in the country. Of the area of Korea, 10% was used as activity data based on systematic sampling, and GHG absorption was estimated applying the factors identified by the IPCC. The result showed GHG absorption by ‘land category’ and ‘urban area’ of 72,990.7 tCO₂·yr^-1 and 201,287.73 tCO₂·yr^-1, respectively. This means that CO₂ absorption can vary depending on the factors that influence the area of vegetation within the settlement. The study identified the share of GHG statistics in settlement sector based on outcomes and contributed to the basis for calculation at the national level.

Measurement of greenhouse gas mitigation is important to developing a National Inventory Report (NIR). However, greenhouse gas inventory has not been conducted for settlement categories, including Land Use, Land Use Change, and Forestry sector (LULUCF), due to the lack of an established methodology. Therefore, this study focused on applying the crown cover method from the Intergovernmental Panel on Climate Change (IPCC)’s Guideline to improve greenhouse gas inventory levels of Tier 2 settlements. Activity data based on crown cover were collected from ortho-images in 2010 and 2017 through supervised classification; these data were classified as the settlement remaining settlement, forest converted to settlement, and the cropland converted to settlement. Estimated crown cover was 493.70, 93.45, and 190.34 ha in 2010, and 656.80, 34.40, and 139.10 ha in 2017 (23.31 ha yr^-1, -8.0 ha yr^-1, and 7.32 ha yr^-1), respectively. CO₂ absorption was calculated as total 247.92 tCO₂ yr^-1, -456.93 tCO₂ yr^-1, and -257.30 tCO₂ yr^-1, respectively. These results had 2.20 - 34.48% variation compared with other studies based on modeling and field surveys. Based on this study, clear spatial definition for settlement is required and activity data, such as national statistics for urban forests, emission factors, and allometric functions using crown cover, should be established.

After the Paris Agreement, the Republic of Korea also has own Nationally Determined Contributions (NDCs), and forest sector has to sequestrate 22 million tons of CO₂ eq. by 2030 and to develop a national inventory and Land‐Use, Land Use Change, and Forestry (LULUCF) system. However, forest activity data have been neglected in the assessment process, because they have been partially recorded from various data sources. Therefore, this study collects available data to represent forest activity data and integrates them considering spatial availability based on Intergovernmental Panel on Climate Change (IPCC) guidelines and approaches from the Kyoto Protocol. National and private spatial data were collected, and the available data as forest activity data were selected. Subsequently, forest management records on address‐based locations were converted to spatial data by spatial joining and overlying. As a result, around 1.25 million forest management and afforestation recodes are available. Based on these records, the estimated average forest management rate considering regional distribution is 31.2% during 2013‐2017, and the estimated forest management rate reflecting average values for missing management recodes is 43.1% during 2003‐2017. This study suggests a spatial approach to develop proper calculation of forest management rate despite differences in forest management rate caused by missing data from filed activities and limitation of spatial data during overlaying analysis. Therefore, more forest data are needed to secure the future development of a national forest inventory.

This study investigated a method for directly estimating methane emissions from rice paddy fields at the field level using drone-based time-series vegetation indices at a town scale. Drone optical and spectral images were captured approximately 15 times from April to November to acquire time-series vegetation indices and optical orthoimages. An empirical regression model validated in previous international studies was applied to calculate cumulative methane emissions throughout the rice cultivation process. Methane emissions were estimated using the vegetation index and yield data were used as input variables for each growth phase. Methane emissions from rice paddies showed maximum values of 309 kg CH₄ ha⁻¹, within a 7% range compared to similar studies, and minimum values of 138 kg CH₄ ha⁻¹, with differences ranging from 29% to 58%. The average emissions were calculated at 247 kg CH₄/ha, revealing slightly lower average values but individual field values within a similar range. The results suggest that drone-based remote sensing technology is an efficient and cost-effective alternative to traditional field measurements for greenhouse gas emission assessments. However, adjustments and validations according to rice varieties and local cultivation environments are necessary. Overcoming these limitations can help establish sustainable agricultural management practices and achieve local greenhouse gas reduction targets.

Detecting forest tree species distribution using satellite imagery with deep-learning models is essential for effective forest management. While sufficient training samples are crucial for developing deep-learning-based tree species classification models, creating these training samples requires significant resources. Therefore, understanding the optimal balance between model accuracy and training sample size is essential for efficient resource allocation. Here, we determined the optimal training sample size for forest tree species classification using Sentinel-2 imagery and the U-Net model. The study area comprised the Seoul–Gyeonggi region of South Korea, where the nine dominant tree species were selected for classification. We utilized multi-temporal Sentinel-2 imagery, incorporating spectral, vegetation, and textural features. Optimal points were identified using Locally Estimated Scatterplot Smoothing (LOESS) regression. The maximum overall accuracy reached 61%, with 90% and 95% of the maximum accuracy with training sample sizes of 2.37%–2.67% and 4.42%–5.89%, respectively. The congeneric Pinus and Quercus groups had major confusion, with species-specific F1-scores ranging from 0.40 (Robinia pseudoacacia) to 0.75 (Pinus koraiensis). These results provide practical guidelines for efficient resource allocation in tree species classification. Rather than pursuing excessive data collection beyond the optimal point, integrating multiple sensor types can overcome existing limitations and enhance classification accuracy.