Harbin, located in northeast China (NEC), has obvious monsoon climate characteristics due to the influence of its geographical environment. Under the control of the polar continental air mass, winter in Harbin is exceedingly cold and long, with the frequent invasion of the cold and dry air from the north. Because of its intensely cold climate in winter, Harbin has created a local form of tourism with its own characteristics: the snow and ice landscape attracts a large number of tourists. Therefore, the anomalies of air temperature and precipitation in winter have an important impact on the livelihood of the local people and economy. In the winter of 2018/2019, the ice and snow tourism in Harbin was harshly affected by the extreme weather, and the direct cause is the anomalies of atmospheric circulation. There is a center of strong positive geopotential height anomalies over east China, which favors the movement of warm air northwards to the NEC, resulting in warmer-than-normal air temperature. Anomalous precipitation is largely controlled by the anomalies of local water vapor and air temperature. The aim of this study was to determine whether the warmer-than-normal temperature, which made the atmosphere more resistant to saturation, was the primary cause of the reduced snowfall. The relative importance of water vapor and air temperature anomalies to the anomalous precipitation was compared. The results suggest that the warmer-than-normal temperature affected all levels, but its impact on the near-surface level was greater. At the middle and upper levels (above 850 hPa), in addition to the warmer-than-normal temperature, the amount of water vapor was less than normal. These conditions both reduced the amount of snow; however, by comparison, the dryness of the air contributed more significantly.

In river and lake ecosystem management, comprehensive water quality monitoring is crucial. Traditional in situ water quality monitoring is costly, and it is challenging to cover entire water bodies. Remote sensing imagery offers the possibility of efficient monitoring of water quality over large areas. However, remote sensing data typically contain a large amount of noise and redundant information, making it difficult for models to capture the effective spectral information and the relationships in the water quality in the remote sensing data. Consequently, this hinders the achievement of high-precision water quality inversion performance. Therefore, this study proposes a comprehensive water quality inversion framework based on a multilayer denoising autoencoder that automatically extracts effective spectral features, utilizing a multilayer denoising autoencoder to extract effective features from Sentinel-2 remote sensing data, thereby reducing noise in the subsequent model input data and mitigating the overfitting problem in subsequent models. A bagging ensemble learning model was established to invert the total phosphorus concentration in Taihu Lake. This model reduces the prediction bias generated by a single machine learning model and was compared with decision tree, random forest, and linear regression models. The research results indicate that compared to a single model, the bagging ensemble learning model achieved better water quality retrieval results, with a coefficient of determination of 0.9 and an MAE of 0.014, while the linear regression model performed the worst, with a coefficient of determination of 0.42. Additionally, models trained using spectral effective information extracted by multilayer denoising autoencoders showed improved water quality retrieval accuracy compared to those trained with raw data, with the coefficient of determination for the bagging model increasing from 0.62 to 0.9. This study provides a rapid and accurate method for large-scale watershed water quality monitoring using remote sensing data, offering technical support for applying remote sensing data to watershed environmental management and water resource protection.

This study analyzed temperature and yield data from 34 meteorological stations in Heilongjiang Province during 1961–2020. Four climate averages (P₁, P₂, P₃, and P₄) were determined based on their respective time distributions (1961–1990, 1971–2000, 1981–2010, and 1991–2020). The national standard temperature anomaly index was used to identify delayed chilling damage in rice cultivation compared to these climate averages. Climate tendency rate analysis, Mann–Kendall detection, and linear regression methods were employed to examine the relationship between temperature anomaly and rice yield from May to September. The results showed that there were noticeable differences in recognizing delayed chilling damage across different climate averages from 1961 to 2020. The average duration of chilling damage under P₁, P₂, P₃, and P₄ was, respectively, estimated as 8.5 years, 13.3 years, 21.4 years, and 30.9 years, with severe cold damage accounting for a significant portion (68.2–76.0%) of the total chilling damage period. The occurrence of severe cold damage increased significantly over time while light and moderate cold damage did not show a clear increasing or decreasing trend. Based on the test results, P₃ was found to be the most suitable climate average for identifying delayed chilling damage in rice cultivation from 1961 to 2020. Moreover, the incidence of chilling damage revealed declining trend over time. There was a high incidence of chilling damage in the 1960s and 1970s, followed by a decrease from the 1980s to the mid 1990s, and finally a low-incidence period after the mid-1990s. Spatially, the western regions experienced greater occurrence of chilling damage than the eastern regions. Additionally, there was a highly significant positive correlation (p < 0.01) between temperature anomalies from May to September and relative meteorological yield of rice. As temperature anomalies decreased during this period, there was an observed downward trend in relative meteorological yield of rice, indicating that delayed cold injury had a negative impact on rice production.

This study analyzed frost formation data provided by the Harbin Meteorological Bureau and considered geographic factors, temperature, and population density. Various analytical methods, including linear fitting, the Mann–Kendall mutation test, the Pettitt method, and the sliding t-test, were employed to identify the temporal and spatial changes as well as the effects of these factors on frost dates in Harbin. The study shows that the first FSD occurred on 18 August, in both 1966 and 1967, which was the 255th day. The latest FSD was observed on 10 October 2006, which was the 283rd day. The earliest occurrence of an FED was on 24 April 2015, which was the 114th day, and the latest was on 21 April 1974, which was the 141st day. The highest number of frost days occurred in 2012, with 161 days, whereas the shortest year was 1966, with only 123 frost days. Throughout the study period, the FSD increased by 7.8 days at a rate of −1.27d/10a, the FED increased by 10.9 days at a rate of 1.77d/10a, and the FFS increased by 18.9 days at a rate of 3.05d/10a. The propensity rates of the FSD and FFS at each location in Harbin indicate an upward trend, while for the FED, certain locations display an upward trend. In general, the FSD has exhibited a delayed trend, the FED has shown an earlier trend, and the FFS has experienced an extended trend. With one-way linear regression, the FSD exhibited an increasing trend at each site, while the FFS also indicated a similar trend, and the FED showed an overall decreasing trend. Throughout the study period, a change was observed in the FSD in 2000, resulting in an average arrival time of the 265th day, or 22 September, of that year. Subsequently, post mutation, the average arrival time of the FSD in the study area was the 272nd day, or 29 September, of that year. In 2006, the FED also underwent a change, with the average arrival time in the study area being the 128th day, or 4 April, of that year. After the change, the average arrival time of the FED in the study area was the 121st day, i.e., 8 April. In 1 April 2004 of that year saw a change in the FFS. Prior to the change, the FFS in the study area averaged the 137th day, whilst following the change, the FFS in the study area averaged the 150th day. The FSD and FFS within Harbin exhibit a negative correlation with latitude and a positive correlation with temperature. Additionally, the FED displays a positive correlation with latitude and a negative correlation with temperature. As the FSD, FED, and FFS in central Harbin are the earliest, latest, and longest, the Pearson correlation coefficient method and multiple regression cannot adequately reflect the effect of longitude.

On 12–13 February 2016, a record-breaking rain–snow event during the passage of an extratropical cyclone occurred in Shandong Province, China, in which the 24 h precipitation totals at 48 of 123 national meteorological stations in Shandong Province broke their historical records for the month of February, and a further 25 stations recorded their second-largest February totals. This paper investigates the evolution of the mesoscale precipitation areas and the mechanisms responsible for the formation, organization, and maintenance of the mesoscale precipitation areas, using FY-2G satellite data, Doppler radar observations, and a Weather Research and Forecasting (WRF) model numerical simulation at 4 km grid spacing. The main results show that the comma head cloud of the cyclone developed from four echo strips. Intense precipitation was related to the mesoscale elongated precipitation areas (EPAs) of reflectivity >30 dBZ within the stratiform clouds. The formation and development of the EPAs coincided with the activities of a low-level shear line and an associated increase in frontogenesis. The simulated EPAs occurred in an environment of conditional instability (CI), inertial instability (II), and conditional symmetric instability (CSI). In the initial stage of the elongated rainfall areas (ERAs), rainfall was initiated by the frontal forcing in the presence of elevated CI, and II was generated by upright convection. During the development stage of the ERAs, the CI was absent, and condensational heating was enhanced. II occurred in the absence of upright convection, and it seems likely that the presence of II is a diabatic signature of the precipitation itself. Upper-level II intensified the convective systems by enhancing outflow aloft, and II caused the ERAs to organize. Thus, II played an important role in the organization and maintenance of the ERAs. The frontogenesis provided the dynamic condition for the release of the instability. Enhanced CSI and II intensified slantwise convection, and combining with enhanced frontogenesis, intensified the ERAs. The echo, ascent, and frontogenesis in snowfall areas were weaker than those in rainfall areas.

In this study, based on frost formation data provided by Harbin Meteorological Bureau, combined with geographical factors, temperature and population density, linear fitting, Mann-Kendall mutation test, pettitt method and sliding T test were used to analyze the spatio-temporal variation characteristics of frost date in Harbin Municipality and the effects of geographical factors, temperature and population density on frost regularity. The results showed that (1) The earliest year of FSD appeared was 1966 and 1967, on the 255th, i.e., August 18th, and the latest was 283 days, i.e., October 10th, 2006. The earliest year in which the FED appeared was April 24th of 2015, the 114th day of that year, and the latest was April 21st of 1974, the 141st day of that year. The year with the most frost was 2012, with 161 days, and the year with the shortest was 1966, with 123 days. (2) Throughout the study period, FSD increased by 7.8 days at a rate of -1.27d/10a, FED increased by 10.9 days at a rate of 1.77d/10a, and FFS increased by 18.9 days at a rate of 3.05d/10a. The tendency rate of FSD and FFS at each site in Harbin showed an increasing trend. For FED, the tendency rate of some sites showed an increasing trend. In general, FSD showed a delayed trend, FED showed an advanced trend, and FFS showed a prolonged trend. Using the method of unary linear regression, FSD of each site showed an increasing trend, FFS showed an increasing trend, and FED showed a decreasing trend in general. (3) Throughout the study period, FSD was mutated in 2000, and the arrival time of it in the study area averaged the 265th, i.e., September 22nd, and after that, the arrival time of it changed to the 272nd, i.e., September 29th of that year. FED was mutated in 2006, and the arrival time of it in the study area averaged 128th, i.e., April 8th. After the mutation, the arrival time of FED changed to 121st, i.e., April 1st of that year. FFS was mutated in 2004. Before the mutation, with an average arrival time of 137days in the study area .But after that, the time changed to 150 days.(4) FSD and FFS in Harbin Municipality were negatively correlated with latitude and positively correlated with temperature, while FED was positively correlated with latitude and negatively correlated with temperature. The FSD, FED and FFS in the central part of Harbin Municipality were the earliest, the latest and the longest, respectively, so the Pearson correlation coefficient method and the influence of longitude could not be reflected in the multiple regression. Several factors were judged by grey correlation analysis method. The longitude is significantly related to FSD, FED and FFS.