In Northern China, many regions and cities are located in semi-arid regions, and groundwater is even the only source of water to support human survival and social development. Affected by human activities, the Bashang (BS) region (including Zhangjiakou City and part of Xilin Gol League) have showed a significant decline in groundwater levels in recent years. However, in the BS region, the causes for the decline in groundwater level were not clear. In this study, we used time series of multi-source data Moderate Resolution Imaging Spectroradiometer (MODIS), Gravity Recovery and Climate Experiment (GRACE) and Global Land Data Assimilation System (GLDAS) to analyze vegetation and groundwater changes based on linear regression models. The variation trends of NDVI (Normalized Difference Vegetation Index, derived from MODIS) and GWSA (groundwater storage anomaly, derived from GRACE and GLDAS) indicated the increasingly better vegetation in the agriculture planting areas, partially degraded vegetation in the grassland, and the declining groundwater level in the whole study region. In order to assess the impact of human-induced factors on vegetation and groundwater, the $RU{E}_{seasonal}$ calculation model was proposed based on RUE (rain use efficiency) in this study. The $RU{E}_{seasonal}$ calculation results showed that human-induced factors promoted the growth of vegetation in agricultural areas and accelerated the consumption of groundwater. In addition, we also obtained temporal and spatial distributions of human activities-affected regions. The area affected by human-induced factors in the south-central study area increased, which accelerated the decline in groundwater levels. From bulletin data, we found that the increasing tourists and vegetable production are respectively the most important factors for the increased consumption of urban water and agricultural water. Based on multi-source data, the influences of various human-induced factors on the ecological environment were explored and the area affected by human-induced factors was estimated. The results provide the valuable guidance for water resource management departments. In the BS region, it is necessary to regulate agricultural water use and strengthen residential water management.

The southern part of the Hebei Province is one of China’s major crop-producing regions. Due to the continuous decline in groundwater level, agricultural water use is facing significant challenges. Precision agricultural irrigation management is undoubtedly an effective way to solve this problem. Based on multisource data (time series soil moisture active passive (SMAP) data, Moderate Resolution Imaging Spectroradiometer (MODIS) normalized difference vegetation index (NDVI) and evapotranspiration (ET), and meteorological station precipitation), the irrigation signal (frequency, timing and area) is detected in the southern part of the Hebei Province. The SMAP data was processed by the 5-point moving average method to reduce the error caused by the uncertainty of the microwave data derived SM. Irrigation signals can be detected by removing the precipitation effect and setting the SM change threshold. Based on the validation results, the overall accuracy of the irrigation signal detection is 77.08%. Simultaneously, considering the spatial resolution limitation of SMAP pixels, the SMAP irrigation area was downscaled using the winter wheat area extracted from MODIS NDVI. The analytical results of 55 winter wheat samples (5 samples in a group) showed that winter wheat covered by one SMAP pixel had an 82.72% growth consistency in surface water irrigation period, which can indicate a downscaling effectiveness. According to the above statistical analysis, this paper considers that although the spatial resolution of SMAP data is insufficient, it can reflect the change of SM more sensitively. In areas where the crop pattern is relatively uniform, the introduction of high-resolution crop pattern distribution can be used not only to detect irrigation signals but also to validate the effectiveness of irrigation signal detection by analyzing crop growth consistency. Therefore, the downscaling results can indicate the true winter wheat irrigation timing, area and frequency in the study area.

This paper proposed a method of extracting the winter wheat area by combining support vector machine (SVM) with variable fuzzy sets. This method mainly aims to deal with mixed pixels in remote sensing images. The SVM classifier can accurately identify pure winter wheat pixels with training samples. However, winter wheat area information in mixed pixels cannot be directly obtained because they contain multiple types of features. In order to estimate the winter wheat area information of mixed pixels in Landsat8 data, this paper introduced the normalized difference vegetation index (NDVI) and the variable fuzzy set method. Finally, this paper reasonably estimated the area of winter wheat from the mixed pixels in Landsat8 images. The accuracy assessment showed that the proposed method could extract winter wheat area information more accurately.

Oil-cellulose insulation systems are widely used in power transformers. Partial discharge (PD) is considered to play an important role in the deterioration process of the oil-cellulose insulation system. There has been little research in the past on the process and mechanism of PD destroying the oil-impregnated paper and pressboard. Through two sets of high voltage experiments on oil-impregnated pressboards, the phenomena about carbonized tracks, and trees were observed and analysed to understand the formation and development mechanism of the carbonized trees inside the pressboards. Furthermore, this analysis was also performed to reveal the mechanism and the process of PDs destroying the pressboards. From the results, we found that PDs occur in the gas channels. The inner wall of the channels is carbonized by the PDs, such that the carbonized tracks and trees propagate along the gas channels. The propagation of the carbonized tree is the process of the thickening and elongating of the tree branches. In addition, in the later stages of the thickening, the energy for carbonization comes from the leakage current and the chemical energy from the combustion of paper.

Oil-cellulose insulation systems are widely used in power transformers. The quality of the insulation system is regarded as a critical aspect of the condition of a transformer. PD activity is a prominent symptom of the ageing and degradation of the insulation system. The results of previous researches are mainly about the white mark generated by PD. The studies of the gas bubble phenomena were not very remarkable. Through three sets of high voltage experiments on oil-impregnated pressboards, we researched and analysed the phenomena of gaseous bubbles and white marks generated by PDs, and the gas channel in the paperboard. The gas composition in the bubbles was obtained. From the results, we found that the intensity of the PD activity has a significant relationship with the bubbles. PDs occur in the gas channels. A major mechanism is that the initial PD decomposes the oil into gases, and the gases become gaseous channels along the tiny pores inside the pressboard.

Accurate monitoring of the spatiotemporal distribution of irrigated areas is crucial for water resource management and food security. Currently, large-scale dynamic irrigation monitoring using remote sensing is typically at the kilometer scale, it does not meet the need for precise monitoring. The growing availability of multi-source optical remote sensing data offers opportunities for field-scale (10–30 m) irrigation monitoring. Multi-source remote sensing synergy improves spatiotemporal continuity but faces challenges in dynamic threshold determination due to irregular data acquisition intervals and sensor retrieval uncertainty. This study proposes a dynamic threshold determination method for high-frequency irrigation detection based on error propagation. First, using Sentinel-2, Landsat-8, GF-1, and GF-6 data, soil moisture time series for the North China Plain from April 10 to 30 are constructed. Second, the soil moisture attenuation patterns and sensor retrieval uncertainties are quantified. Finally, derived dynamic irrigation identification thresholds (T) based on observation intervals (D) and sensor combinations. Results show that Sentinel-2 has the highest retrieval stability (standard deviation of 4.24%), while GF-1 has the largest uncertainty (4.79%).The overall accuracy of irrigation event identification reaches 92.5%. Comparative experiments demonstrate that this method outperforms multi-year average and change detection methods, effectively reducing missed detections due to long observation intervals and mitigating misidentifications caused by neglecting soil moisture attenuation. This method provides technical support for high-frequency irrigation monitoring at the field scale and dynamic estimation of regional irrigation water use.