Satellite optical-infrared remote sensing from the Moderate Resolution Imaging Spectroradiometer (MODIS) provides effective air temperature (T_a) retrieval at a spatial resolution of 5 km. However, frequent cloud cover can result in substantial signal loss and remote sensing retrieval error in MODIS T_a. We presented a simple pixel-wise empirical regression method combining synergistic information from MODIS T_a and 37 GHz frequency brightness temperature (T_b) retrievals from the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E) for estimating surface level T_a under both clear and cloudy sky conditions in the United States for 2006. The instantaneous T_a retrievals showed favorable agreement with in situ air temperature records from 40 AmeriFlux tower sites; mean R² correspondence was 86.5 and 82.7 percent, while root mean square errors (RMSE) for the T_a retrievals were 4.58 K and 4.99 K for clear and cloudy sky conditions, respectively. Daily mean T_a was estimated using the instantaneous T_a retrievals from day/night overpasses, and showed favorable agreement with local tower measurements (R² = 0.88; RMSE = 3.48 K). The results of this study indicate that the combination of MODIS and AMSR-E sensor data can produce T_a retrievals with reasonable accuracy and relatively fine spatial resolution (~5 km) for clear and cloudy sky conditions.

Many process-based models for carbon flux predictions have faced a wide range of uncertainty issues. The complex interactions between the atmosphere and the forest ecosystems can lead to uncertainties in the model result. On the other hand, artificial intelligence (AI) techniques, which are novel methods to resolve complex and nonlinear problems, have shown a possibility for forest ecological applications. This study is the first step to present an objective comparison between multiple AI models for the daily forest gross primary productivity (GPP) prediction using satellite remote sensing data. We built the AI models such as support vector machine (SVM), random forest (RF), artificial neural network (ANN), and deep neural network (DNN) using in-situ observations from an eddy covariance (EC) flux tower and satellite remote sensing data such as albedo, aerosol, temperature, and vegetation index. We focused on the Gwangneung site from the Korea Regional Flux Network (KoFlux) in South Korea, 2006–2015. As a result, the DNN model outperformed the other three models through an intensive hyperparameter optimization, with the correlation coefficient (CC) of 0.93 and the mean absolute error (MAE) of 0.68 g m⁻² d⁻¹ in a 10-fold blind test. We showed that the DNN model also performed well under conditions of cold waves, heavy rain, and an autumnal heatwave. As future work, a comprehensive comparison with the result of process-based models will be necessary using a more extensive EC database from various forest ecosystems.

Forest fire modeling often requires estimates of fuel moisture status. Among the various fuel variables used for fire modeling studies, the 10-h fuel moisture content (10-h FMC) is a promising predictor since it can be automatically measured in real time at study sites, yielding more information for fire models. Here, the performance of 10-h FMC models based on three different approaches, including regression (M_REG), machine learning algorithms (M_ML) with random forest and support vector machine, and a process-based model (M_FSMM), were compared. In addition, whole-year models of each type were compared with their respective seasonal models to explore whether the development of separate seasonal models yielded better estimates. Meteorological conditions and 10-h FMC were measured each minute for 18 months in and near a forest site and used for constructing and examining the 10-h FMC models. In the assessments, M_ML showed the best performance (R² = 0.77–0.82 and root mean squared error [RMSE] = 2.05–2.84%). The introduction of the correction coefficient into M_REG improved its estimates (R² improved from 0.56–0.58 to 0.68–0.70 and RMSE improved from 3.13–3.85% to 2.64–3.27%) by reducing the errors associated with high 10-h FMC values. M_FSMM showed the worst performance (R² = 0.41–0.43 and RMSE = 3.70–4.39%), which could possibly be attributed to the lack of radiation input from the study sites as well as the particular fuel moisture stick sensor that was used. Whole-year models and seasonal models showed almost equal performance because 10-h FMC varied in response to atmospheric moisture conditions rather than specific seasonal patterns. The adoption of a hybrid modeling approach that blends machine-learning and process-based approaches may yield better predictability and interpretability. This study provides additional evidence of the lagged response of 10-h FMC after rainfall, and suggests a new way of accounting for this response in a regression model. Our approach using comparisons among models can be utilized for other fire modeling studies, including those involving fire danger ratings.

Leaf area index is a key structural parameter for biological and physical processes. Korea is planning to launch CAS500-4 in 2025, so in situ data is needed to validate the leaf area index. Unlike other networks (e.g., NEON and TERN), establishing an elementary sampling unit is difficult in Korea due to the complex forest structure and rugged terrain. Therefore, pixel-level correspondence between the satellite product and fisheye footprints is the best way to verify in complex terrain. In this study, we analyzed the spatial footprint of fisheye lenses in different forest types using terrestrial LiDAR data for the first time. The three-dimensional forest structure was analyzed at various viewing zenith angles, and the footprint radius was approximately 3 m at view zenith angle (VZA) 20° and approximately 10 m at VZA 90°. We also analyzed the Z-values from terrestrial laser data and the plant area index on leafless seasons to assess the impact of obstacles, such as tree trunks, under various viewing zenith angles. The analysis showed that the influence of woody components increases dramatically as the VZA exceeds 40°. Such factors influenced the increase in LAI and the decrease in the clumping index as the VZA increased. Overall, we concluded that narrowing VZA between 20° and 40° is appropriate for Korean forests with complex structures.

Droughts affect economic, social, and environmental aspects in regions such as the Korean Peninsula, where more than 70% of the area comprises forests; hence, their monitoring is imperative. Despite the many indices and methodologies developed for monitoring, diagnosing droughts using reanalysis data is challenging as the data are characterized by low resolution and simplified vegetation classification. This study utilized a recently released ERA5 reanalysis dataset and its vegetation type information to derive indices that represent meteorological drought. Furthermore, their accuracy in South Korea, based on observation data, was evaluated. The spatio-temporal variability of droughts was analyzed using various factor and correlation analysis methods considering different atmospheric variables and soil moisture. The validity of the method was verified by comparing the observed and obtained data. Soil moisture in the first and second soil layers was sensitive to droughts in low-vegetation areas, thus requiring relatively frequent monitoring of precipitation and evapotranspiration near the topsoil. High-vegetation areas were most affected in the third layers one month after the drought. Hence, forest drought monitoring should consider precipitation, evapotranspiration, and runoff one month in advance. The results obtained herein can be used for forest drought monitoring one month before its occurrence.

Recent trends of intense forest fires in the Korean Peninsula have increased concerns about more extreme burning in the future under a warming climate. Accurate and reliable fire weather information has become more critical to reduce the risk of forest-related disasters over complex terrain. In this study, two parameterizations reflecting complex topographic effects were implemented in the Weather Research and Forecasting (WRF) model. The model performance was evaluated over the mountainous region in Gangwon-do, South Korea’s most significant forest area. The simulation results of the wildfire case in 2019 show that subgrid-scale orographic parameterization considerably improves model performance regarding wind speed, with a lower root mean square error (RMSE) and bias by 53% and 57%, respectively. Another parameterization, reflecting slope and shading, effectively reflected sunrise and sunset effects. The second parametrization produced little effect on the daily averages of meteorological elements. However, thermodynamic components such as temperature and heat flux show more realistic values during sunset or sunrise when the solar altitude angle is low. The results imply that applying topographic parameterizations is required in numerical simulations, especially for hazardous weather conditions over complex terrain in mountainous regions.