Lithology classification is important in mineral resource exploration, engineering geological exploration, and disaster monitoring. Traditional laboratory methods for the qualitative analysis of rocks are limited by sampling conditions and analytical techniques, resulting in high costs, low efficiency, and the inability to quickly obtain large-scale geological information. Hyperspectral remote sensing technology can classify and identify lithology using the spectral characteristics of rock, and is characterized by fast detection, large coverage area, and environmental friendliness, which provide the application potential for lithological mapping at a large regional scale. In this study, ZY1-02D hyperspectral images were used as data sources to construct a new two-layer extreme gradient boosting (XGBoost) lithology classification model based on the XGBoost decision tree and an improved greedy search algorithm. A total of 153 spectral bands of the preprocessed hyperspectral images were input into the first layer of the XGBoost model. Based on the tree traversal structural characteristics of the leaf nodes in the XGBoost model, three built-in XGBoost importance indexes were split and combined. The improved greedy search algorithm was used to extract the spectral band variables, which were imported into the second layer of the XGBoost model, and the bat algorithm was used to optimize the modeling parameters of XGBoost. The extraction model of rock classification information was constructed, and the classification map of regional surface rock types was drawn. Field verification was performed for the two-layer XGBoost rock classification model, and its accuracy and reliability were evaluated based on four indexes, namely, accuracy, precision, recall, and F1 score. The results showed that the two-layer XGBoost model had a good lithological classification effect, robustness, and adaptability to small sample datasets. Compared with the traditional machine learning model, the two-layer XGBoost model shows superior performance. The accuracy, precision, recall, and F1 score of the verification set were 0.8343, 0.8406, 0.8350, and 0.8157, respectively. The variable extraction ability of the constructed two-layer XGBoost model was significantly improved. Compared with traditional feature selection methods, the GREED-GFC method, when applied to the two-layer XGBoost model, contributes to more stable rock classification performance and higher lithology prediction accuracy, and the smallest number of extracted features. The lithological distribution information identified by the model was in good agreement with the lithology information verified in the field.

Fast and accurate SOM estimation and spatial mapping are significant for cultivated land planning and management, crop growth monitoring, and soil carbon pool estimation. It is a key problem to construct a fast and efficient estimation model based on hyperspectral remote sensing image data to realize the inversion mapping of SOM in large areas. In order to solve the problem that the estimation accuracy is not high due to the influence of hyperspectral image quality and soil sample quantity during the estimation model construction, this study explored a method for constructing an estimation model of SOM contents based on a new stacking ensemble learning algorithm and hyperspectral images. Surface soil samples in Huangzhong County of Qinghai Province were collected, and their ZY1-02D hyperspectral remote sensing images were investigated. As input data, a feature band dataset was constructed using the Pearson correlation coefficient and successive projections algorithm. Based on the dataset, a new SOM estimation model under the stacking ensemble learning framework combined with heterogeneous models was developed by optimizing the combination of base and meta-learners. Finally, the spatial distribution map of SOM was plotted based on the result of the model over the study area. The result suggested that the input data quality of the estimation model is improved by constructing a feature band dataset. The multi-class ensemble learning estimation model with the combination strategy of the base and meta-learners has better predictive effects and stability than the single-algorithm and single-level ensemble models with homogeneous learners. The coefficient of determination is 0.829, the residual prediction deviation is 2.85, and the predictive set root mean square error is 1.953. The results can provide new ideas for estimating SOM content using hyperspectral images and ensemble learning algorithms, and serve as a reference for mapping large-scale SOM spatial distribution using space-borne hyperspectral images.

Net primary productivity (NPP) of vegetation is considered an important indicator for ecological stability and is the main object for analyzing the factors influencing the terrestrial carbon cycle. Recent studies have made clear the changes in the NPP of vegetation and its influencing factors at various scales. However, the variations in NPP based on different land cover types under various natural conditions, along with their driving factors, remain not well understood. In this study, spatial overlay analysis was used to investigate the link among climatic, soil moisture (SM), and topographic parameters and NPP of various land cover types after analyzing the spatial and temporal trends of NPP in the Songnen Plain from 2001 to 2020. Additionally, the contribution of each influence factor to the NPP of different land cover types was calculated using the elastic net regression model. The elastic net regression model eliminates the multicollinearity among the influencing factors while maintaining the model stability, and the R² of all lands is greater than 0.62, which can effectively quantify the contribution of each influencing factor to NPP. The results show a continuously increasing trend of the overall NPP in the research area over the selected 20 years, and NPP increased most significantly in forest land (FOR). Precipitation (PRE) and NPP showed high correlations in all the different land cover types, while the correlations between NPP and other influencing factors were significantly different. In addition, we found that perennials led to a more significant degree of NPP enhancement, and the effect of topographic conditions on NPP was mainly reflected in differences in moisture conditions due to surface runoff. From the results of the modeling calculations, the cumulative contribution of PRE to NPP ranks first in all land types and is the most vital influencing factor of NPP in the Songnen Plain. SM was an important influence, but the contribution of NPP was greater in land classes with shallow root systems. The results of the study revealed the positive transformation relationship of NPP among land cover types in ecologically fragile areas, which provides a reference for ecological restoration and rationalization of land use structure in zones such as intertwined agricultural and pastoral zones.

Evapotranspiration (ET) is a vital constituent of the hydrologic cycle. Researching changes in ET is necessary for understanding variability in the hydrologic cycle. Although some studies have clarified the changes and influencing factors of ET on a regional or global scale, these variables are still unclear for different land cover types due to the range of possible water evaporation mechanisms and conditions. In this study, we first investigated spatiotemporal trends of ET in different land cover types in the Xiliao River Plain from 2000 to 2019. The correlation between meteorological, NDVI, groundwater depth, and topographic factors and ET was compared through spatial superposition analysis. We then applied the ridge regression model to calculate the contribution rate of each influencing factor to ET for different land cover types. The results revealed that ET in the Xiliao River Plain has shown a continuously increasing trend, most significantly in cropland (CRO). The correlation between ET and influencing factors differed considerably for different land cover types, even showing an opposite result between regions with and without vegetation. Only precipitation (PRCP) and NDVI had a positive impact on ET in all land cover types. In addition, we found that vegetation can deepen the limited depth of land absorbing groundwater, and the influence of topographic conditions may be mainly reflected in the water condition difference caused by surface runoff. The ridge regression model eliminates multicollinearity among influencing factors; R² in all land cover types was over 0.6, indicating that it could be used to effectively quantify the contribution of various influencing factors to ET. According to the results of our model calculations, NDVI had the greatest impact on ET in grass (GRA), cropland (CRO), paddy (PAD), forest (FOR), and swamp (SWA), while PRCP was the main influencing factor in bare land (BAR) and sand (SAN). These findings imply that we should apply targeted measures for water resources management in different land cover types. This study emphasizes the importance of comprehensively considering differences among various hydrologic cycles according to land cover type in order to assess the contributions of influencing factors to ET.