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.

We quantified the extent and severity of Asian gypsy moth (Lymantria dispar) defoliation in Wonju, Korea, from May to early June in 2020. Landsat images were collected covering Wonju and the surrounding area in June from 2017 to 2020. Forest damage was evaluated based on differences between the Normalized Difference Moisture Index (NDMI) from images acquired in 8 June 2020 and the prior mean NDMI estimated from images in June from 2017 to 2019. The values of NDMI ranged from −1 to 1, where values closer to 1 meant higher canopy cover. The NDMI values for 7825 ha of forests were reduced by more than 0.05 compared to the mean NDMI values for the prior 3 years (2017 to 2019). The NDMI values of 1350 ha of forests were reduced by >0.125 to 0.2, and the NDMI values for another 656 ha were reduced by more than 0.2. A field survey showed that these forests were defoliated by gypsy moth and that forests with NDMI reductions of more than 0.2 were heavily defoliated by gypsy moth. A 311 ha area of Japanese larch (Larix kaempferi) was severely damaged by gypsy moth and the proportion of larch damaged was higher than that of other tree species. This intense damage to larch suggests that gypsy moths preferentially attack Japanese larch in Wonju. Our study shows that the use of NDMI values to detect areas defoliated by gypsy moth from satellite images is effective and can be used to measure other characteristics of gypsy moth defoliation events, such as host preferences under field conditions.

The role of forests is increasing because of rapid land use changes worldwide that have implications on ecosystems and the carbon cycle. Therefore, it is necessary to obtain accurate information about forests and build forest inventories. However, it is difficult to assess the internal structure of the forest through 2D remote sensing techniques and fieldwork. In this aspect, we proposed a method for estimating the vertical structure of forests based on full-waveform light detection and ranging (FW LiDAR) data in this study. Voxel-based tree point density maps were generated by estimating the number of canopy height points in each voxel grid from the raster digital terrain model (DTM) and canopy height points after pre-processing the LiDAR point clouds. We applied an unsupervised classification algorithm to the voxel-based tree point density maps and identified seven classes by profile pattern analysis for the forest vertical types. The classification accuracy was found to be 72.73% from the validation from 11 field investigation sites, which was additionally confirmed through comparative analysis with aerial images. Based on this pre-classification reference map, which is assumed to be ground truths, the deep neural network (DNN) model was finally applied to perform the final classification. As a result of accuracy assessment, it showed accuracy of 92.72% with a good performance. These results demonstrate the potential of vertical structure estimation for extensive forests using FW LiDAR data and that the distinction between one-storied and two-storied forests can be clearly represented. This technique is expected to contribute to efficient and effective management of forests based on accurate information derived from the proposed method.

It is crucial to evaluate the effects of thinning on litterfall production, soil chemical properties, and fine root dynamics when implementing thinning as a silvilcultural technique to enhance tree growth and timber yield in Pinus koraiensis plantations. Thus, we determined the 10-year effects (2007–2017) of different thinning intensities on litterfall production, soil chemical properties, and fine root biomass and necromass within a P. koraiensis plantation in South Korea. The soil chemical parameters and fine root biomass and necromass were also compared across three soil depths (0–10, 10–20, and 20–30 cm). Three thinning treatments were employed: no thinning (CON), light thinning (32% removed, LT), and heavy thinning (64% removed, HT). Results revealed that litterfall was consistent across all thinning treatments, but broadleaf species had considerably higher litterfall production at HT stands than at CON/LT stands. Soil chemical properties, except exchangeable K⁺, were generally lower at LT stands, particularly at a depth of 20–30 cm soil. After ten years, there was a decrease in fine root biomass and necromass with increasing soil depth. Over 80% of fine roots were found in the upper layer (0–20 cm), while very fine roots (0–1 mm) consisted mainly of 47% pine and 53% other species and were concentrated in the 0–10 cm soil depth in HT. In conclusion, different thinning intensities had diverse effects on the parameters measured within the plantation. Future studies can explore how the effects of thinning intensities on litterfall production, soil chemistry, and fine root dynamics affect species diversity, carbon storage, and understory vegetation in P. koraiensis plantations.

Forest type maps require periodic revision, yet field inspection of every stand is impractical. This study tested whether stand-level disagreement between a Sentinel-2 U-Net classifier and the Republic of Korea’s forest type map could support screening for revision. A seven-class U-Net was trained using seasonal Sentinel-2 composites from 2025, and pixel predictions were aggregated to forest stands. Disagreeing stands passing dominance and confidence thresholds were flagged and assessed by blinded interpreters using high-resolution basemaps supplemented by Sentinel-2 composites. Stand-level agreement with the map was 0.69 by stand count and 0.76 when area-weighted. In the primary analysis, high-precision-class flags were mismatched in 41.3% of interpreted stands (29.7% when population-weighted), compared with 7.7% of controls. Test-set class precision, the pre-specified prioritization criterion, did not distinguish more reliable flags in this analysis. An exploratory partition by flag type yielded weighted rates of corresponding mismatches of 80.9% for stocking-status flags and 26.0% for species-dominance flags, representing 12.3- and 2.4-fold enrichment over their respective region-wide baselines. Approximately 35% of flagged stands were estimated to be mismatches, and the flags captured approximately 7.5% of estimated mismatches. These findings support model–map disagreement as a potential tool for prioritizing stand verification in the study region.

The present study aimed to date the death of Korean red pine (Pinus densiflora Siebold & Zucc.) at Sogwang-ri, where is famous for the Korean red pine forests in Korea, using a dendrochronological method, and to investigate the effect of climatic factors on the death and growth decline of the pine trees. For the study, 51 Korean red pines were selected from three stands (Jungmi-gol A, Jungmi-gol B, and Neonam-gol) in the study area, and for the ring-width measurement the increment cores were extracted from the selected trees. The stand chronologies for the Jungmi-gol and Neonam-gol were established using the mean values of individual time series of the living trees from each site, and the chronologies were further used for cross-dating with individual time series of the dead trees to date their death. The abrupt growth reduction was also investigated based on the consecutive ring-width growth decline trend to know the vitality during the living. Results revealed that the red pines at Jungmi-gol A were dead in 2020, at Jungmi-gol B in 2012 or 2013, and at Neonam-gol in 2019. So, it was verified that the death of the red pines at the study sites occurred from high to low elevations. In the analysis of the role of climate in the annual ring growth, the trees at Jungmi-gol A and B located on the southwest slope showed more ring-width growth when precipitation in April, July, and August were increasing, at Neoman-gol located on the north slope when precipitation only in July was increasing. Although the correlations of the ring-width growths at the sites with temperature and precipitation were somehow different from each other, Drought Stress Index (DSI) from May to September at all sites were all negative. Therefore, water stress caused by increased temperature and decreased precipitation is the most significant climatic parameter for the death or growth decline of the pine trees in the study area.