Assessing and predicting forest fires has long been an arduous task. Nowadays, the rapid advancement of artificial intelligence and machine learning technologies have provided a novel solution to forest fire occurrence assessment and prediction. In this research, we developed a novel hybrid machine-learning-technique algorithm to improve forest fire prediction based on random forest (RF), gradient-boosting decision tree (GBDT), support vector machine (SVM), and other machine learning models. The dataset we employed was satellite fire point data from 2010 to 2018 from the Chinese Department of Fire Prevention. The efficacy and performance of our methods were examined by validating the model fit and predictive capability. The results showed that the ensemble model LR (logistic regression)-RF-SVM-GBDT outperformed the single RFSVMGBDT model and the LR-RF-GBDT integrated framework, displaying higher accuracy and greater robustness. We believe that our newly developed hybrid machine-learning algorithm has the potential to improve the accuracy of predicting forest fire occurrences, thus enabling more efficient firefighting efforts and saving time and resources.

Understanding the vertical structures of cloud condensation nuclei (CCN) and ice-nucleating particle (INP) number concentrations in desert source regions is crucial for examining dust-cloud interactions and other related impacts. To explore the vertical profiles of the CCN and INP number concentrations and their possible atmospheric–dynamic influence factors at the center of the Taklimakan Desert, intensive observations were conducted by employing a ground-based polarization Raman lidar, sounding balloons, and a sun photometer in Tazhong (83.39° E, 38.58° N, 1103 m above sea level) during the summer of 2019. Based on the GRASP algorithm, the extinction-to-volume conversion factor of dust aerosols was 0.85 × 10⁻¹² Mmm³ m⁻³, and the extinction-to-number conversion factor was predicted to be 0.20 Mm cm⁻³ on the basis of the sun photometer observations. Thus, the vertical CCN and INP number concentration profiles obtained with different parameterization schemes in the presence of various pollution levels were calculated by combining dust extinction coefficients retrieved by lidar and meteorological data observed by sounding balloon observations. The achieved results indicated that the CCN number concentration varied from 10⁻² to 10² cm⁻³ and decreased from ground level to 12 km with an average value of 36.57 cm⁻³ at the 10–12 km height range, while the INP number concentration based on parameterization schemes D10 and D15 mainly varied from 10⁻¹ to 10² L⁻¹ and from 1 L⁻¹ to 10³ L⁻¹, with average values of 3.50 L⁻¹ and 7.80 L⁻¹, respectively. Moreover, we observed a strong relationship between the INP number concentration of scheme D10 and the wind speed, with an R² value of 0.72, but a weak relationship between the CCN number concentration and the relative humidity in the boundary layer, with a Spearman’s rank correlation coefficient R² value of 0.38. The present study provides original and valuable information regarding the CCN and INP number concentrations and their related influencing factors at the center of the Taklimakan Desert and can improve our understanding of the vertical distributions of dust–cloud–atmosphere dynamic interactions, as well as of the roles of dust aerosols in the desert hydrological cycle.

In current forestry investigation studies, the research hotspots have tended to concentrate on ascertaining the precision of certain tree parameters. This has resulted in an augmented intricacy of the technique in terms of algorithms and observation instruments. The complexity of the technology and the cost of the equipment make it impossible to use for large-scale forest surveys, for example, a national forest inventory (NFI). The aim of our study was to design a new type of low-cost measuring method that could be utilized in a NFI and in developing countries. Meanwhile, the newly designed method was expected to be able to output certain forest measurement factors without necessitating data processing by NFI field investigators. Based on these objectives, we developed a measuring method that included hardware comprised of two tools. The first tool was an electronic measuring tape that contained a microcontroller unit (MCU) and could automatically record and collaborate with other equipment via wireless protocols. The second tool was a tree stem position mapper that utilized our own designed mechanisms. The results showed that the tree DBH measurements exhibited a 0.05 cm (0.20%) bias and a 0.36 cm (1.45%) root mean square error (RMSE), and the biases on the x-axis and the y-axis of the tree position estimations were −15.92–9.92 cm and −25.90–10.88 cm, respectively, accompanied by corresponding RMSEs of 15.27–29.40 cm and 14.49–34.68 cm. Moreover, an efficiency test determined that the average measurement time per tree was 20.34 s, thus, demonstrating a marked improvement in speed by nearly one-fold compared to the conventional method. Meanwhile, this measurement kit costs less than 150 Euros and is economically suitable for large-scale applications. We posit that our method has the potential to serve as a standard tool in a Chinese NFI and in developing countries in the future.

Land–atmosphere interactions are influenced by the earth’s complex underlying subsurface, which in turn indirectly affects atmospheric motion and climate change. Human activities are increasingly exerting an influence on desert ecosystems, and artificial green land with clear functional orientation has been established in many desert areas. Consequently, the previously dominant, shifting, sand-covered, underlying surface in these desert regions is gradually transforming. This transformation has significant implications for the characteristics of land–atmosphere interactions, causing them to deviate from their original state. At present, existing studies still have not presented a systematic understanding of this change and have ignored the impact of human activities on land–atmosphere interactions in artificial green land. To address these research gaps, this study specifically targets artificial green land in the Tazhong region of Taklamakan Desert. We carried out observation experiments on land–atmosphere interactions in three different functional units from outside to inside: natural shifting sands, the shelter forest, and the living area. We also analyzed the differences and attribution of land–atmosphere interactions characteristics of different functional units. Compared with the natural shifting sands, the daily average maximum values of wind speed in the shelter forest decreased by 78%, and the daily average maximum air temperature and soil (0 cm) temperature decreased by 2.6 °C and 7 °C, respectively. Additionally, the soil moisture level was significantly increased throughout the green land due to the shelter forest. The surface albedo experienced a decrease, with an annual average of 0.21. Furthermore, the aerodynamic roughness and bulk transport coefficient increased by two orders of magnitude. The daily average maximum values of sensible heat flux and soil heat flux (G₀₅) decreased by 18.7% and 75%, respectively, and the daily average maximum value of latent heat flux increased by 70.3%. This effectively improved the microclimate environment of the green land. The living area was greatly reduced by the shelter forest coverage and influenced by the buildings. Consequently, the environmental improvement was not as large as it was inside the shelter forest. However, it still provided a good shelter for production and living in the desert area. Throughout the year, a total of 4.60 × 10⁵ t water was consumed through evapotranspiration in the artificial green land. The findings of this study have the potential to enhance our comprehension of land–atmosphere interactions in desert regions, thereby offering valuable insights for the establishment and effective management of artificial desert green lands.

Continuous monitoring of evapotranspiration (ET) at high spatio-temporal resolutions is vital for managing agricultural water resources in arid and semi-arid regions. This study used the enhanced spatial and temporal adaptive reflectance fusion model (ESTARFM) to calculate the ET of winter wheat between the green-up and milk stages in Linfen Basin, a typical, semi-arid area of the Loess Plateau, at temporal and spatial resolutions of 30 m and 8 d, respectively. We then analyzed the impact of meteorological factors on ET and its variation during the main growth period of winter wheat. The fused ET data displayed the spatial details of the OLI ET data better and could accurately reflect ET variation and local sudden variations during the main growth period of winter wheat. Moreover, winter wheat ET in rain-fed areas is more heavily influenced by meteorological factors, and the effect is more direct. Affected by the synergistic effect of wind velocity, precipitation, and temperature, the ET of winter wheat in rain-fed area was lower in the green-up stage. Then, ET gradually increased, reaching its maximum in the heading–grain filling stage. At the jointing stage, temperature had a significant effect on ET. A combination of precipitation and temperature had the greatest impact on the ET of winter wheat in the heading–filling stage. In the milk stage, meteorological factors had a minor impact on ET. This study serves as a reference for ET in winter wheat in semi-arid areas and its influencing meteorological factors, which can assist in drought mitigation and regional food security strategies.

Carbon neutrality is an important target in China’s efforts to combat the climate crisis. The implementation of carbon neutrality requires high crop yields in farmland ecosystems of arid regions. However, the responses of farmland ecosystems to environmental changes and their effects on the conversion and intensity of carbon sources/sinks within farmlands in arid regions remain unclear, which limits carbon sequestration. In this study, we used a set of eddy covariance systems to observe carbon and water fluxes in cotton and spring maize, two typical crops in arid regions of Northern Xinjiang in China. The carbon and water exchange and water use efficiency (WUE) of cotton and spring maize were evaluated over the entire growth cycle with respect to changes in the environment. Our results show that the carbon sequestration capacity of farmland ecosystems in arid regions is undeniable and is strongly influenced by the growth and development of plants. Spring maize, as a representative of C4 plants, exhibited a 58.4% higher carbon sequestration efficiency than cotton, a C3 plant, and they both reached their carbon sequestration efficiency peak in July. Throughout the growth period, temperature, net surface radiation, and saturated vapor pressure differences (VPD) significantly affected the carbon sequestration capacity and WUE of both crops. Optimal temperatures can maximize the carbon sequestration efficiency of cotton and spring maize; for cotton, they are 20–25 °C, and for spring maize, they are 22–27 °C, respectively. In addition, it is recommended that spring maize be harvested at the end of July when it meets the harvesting standards for silage feed and achieves the maximum carbon sequestration. Afterward, winter crops should be planted to maximize the yield and improve the carbon sequestration capacity of farmlands.