The NCEP FNL (National Centers for Environmental Prediction Final Operational Global Analysis data) reanalysis data was used to diagnose the heavy snowfall process in the Liaoning region from February 14 to 16, 2020, and analyze the forecast deviation of the numerical model and its correlation with the affecting system.The results show that, under the influence of the upper trough in North China and the low overlying trough in the north of the Yellow Sea, the warm and humid air in the southern branch front area climbes along the cold pad of the northern wind in the northern branch front area, forming water vapor convergence and uplifting movement, which provides favorable dynamic conditions for the occurrence of the heavy snowfall process.The performance of the CMA-GFS (China Meteorological Administration Global Forecast System) model is poor in predicting the precipitation intensity of the heavy snowfall center, but the prediction of the heavy snowfall area is relatively accurate.The ECMWF (European Center for Medium-Range Weather Forecasts) model is more accurate in predicting precipitation intensity, but the forecast of heavy snowfall area is biased westward and the range looks too larger.The ECMWF model is relatively accurate in predicting the long duration (3.5 d) of this heavy snowfall process, and the precipitation forecast has a turning change and obvious forecasting biases in the short duration.The diagnostic results show that the turning change of precipitation forecast during the heavy snowfall is significantly correlated with the forecast deviation of the high and low altitude low-pressure systems.In particular, the forecast deviation of the low altitude low-pressure system is significantly correlated with the specific humidity of the two water vapor channels over the heavy snowfall area.The stronger forecast of the low-pressure system corresponds to the larger specific humidity, resulting in stronger water vapor transport over the Sea of Japan in the model.It may be the cause of the obvious deviation in the precipitation forecast.

A typical strong sand-dust event in Northern China from March 27 to 28, 2021 was analyzed using conventional meteorological data, sounding station meteorological observation data, retrieval data of sand-dust from satellite cloud images and the NCEP/NCAR (National Center for Environmental Prediction/National Center for Atmospheric Research) reanalysis data.The results show that the dry (the average precipitation is less than 10 mm, and the area with negative precipitation anomaly is more than 80%) and hot (the temperature is 6℃ higher than usual) weather are the climatic background factors for the dust weather.At the same time, under the influence of the Mongolian cyclone, the strong wind at medium-low altitude and on the ground, the strong vertical ascending motion is the dynamic conditions for the dust weather.Based on the atmospheric temperature lapse rate, uplift index, K index, convective effective potential energy index, I_Conve index from 700 hPa to 850 hPa, I_Conve index from 700 hPa to the ground, and PW index, it indicates that the atmospheric structure in the sand source area is unstable, which provides thermal conditions for the occurrence of dust weather.The 24-hour backward trajectory analysis shows that the dust near the surface of Inner Mongolia is uplifted to 1000 m and transported over a long distance to a low level of about 500 m at Shenyang.The dust at the height of 4000~5000 m is transported to the high level about 1000~1500 m at Shenyang.

Using conventional meteorological data, particulate matter monitoring and laser radar detection data, the causes and transmission characteristics of a severe sandstorm weather event in Hohhot in March 2021 were analyzed by methods of weather analysis, and air quality trajectory tracking models.The results showed that there were two peaks of pollutant concentration during this pollution process, which can be divided into two stages.From March 14 to 15, under the combined influence of an upper-level trough and a surface cold front, Hohhot experienced strong winds, blowing sand, sandstorms, and severe sandstorms.The PM_2.5 rose to 1382 mg·m^-3 at 03:00 on the 15th, and the duration of severe pollution reached 15 h.From March 16 to 18, under the weak west-ridge in the upper level and weakened surface pressure field, the near-surface diffusion conditions were relatively poor, resulting in the accumulation and return of pollutants and the prolonged duration of severe pollution.Laser radar data showed that during the daytime of March 14, there were a large number of fine particles gathered in the upper air over Hohhot.In the early morning of the 15th, the primary pollutant in this area changed from PM_2.5 to PM₁₀.At 12:00, the PM₁₀ concentration rose to 4099 mg·m^-3.On the 16th, dust returned, causing severe pollution again.On the 17th, the pollutant concentration began to decrease, and the dust weather ended at 05:00 on the 18th.Backward trajectory analysis showed that Mongolia was the main source area for this severe dust pollution.

From November 17 to 20, 2020, there was a large-scale freezing rain and snow weather process in Jilin province, and there was a long-term and large-scale freezing rain during the transition of rain and snow, causing a rare degree of disaster in history.In this paper, the conventional ground observation data, radiosonde second data, microwave radiometer data, and ERA5 reanalysis data were used to analyze the cause mechanism of the freezing rain.The results show that the process is influenced by the deepening of the eastward movement of the upper trough, the southward invasion of the Mongolian High, and the northward surface cyclone.The low-level jet brings abundant water vapor to the precipitation in Jilin province, and the specific humidity is generally 3~7 g·kg^-1.At the same time, under the influence of a warm ridge at 850 hPa, the 0℃ isotherm maintains and moves slowly in the central part of Jilin province for a long time, and the cold and warm air masses confront each other for a long time, forming a stable inversion layer in the middle and lower troposphere, which provides favorable temperature stratification conditions for freezing rain in Changchun city and Jilin city.Through the analysis of the temperature and humidity stratification during freezing rain in Changchun and Jilin stations, it is found that the freezing rain occurs in the cold-warm-cold temperature stratification from the bottom up of the vertical temperature field.The height of the cold pad is higher than 950 hPa, the thickness is 0.7~1.0 km, the temperature at the lower 1000 hPa is -3~0℃, the temperature at 850 hPa is 1~3℃, and the thickness of the warm layer is at 950~750 hPa, about 1.5 km.The maximum temperature in the center of the warm layer is greater than 3℃.The content of solid water in the middle and upper altitudes (800~350 hPa) should not be less than 0.2 g·kg^-1, and the content of liquid water in the middle and lower altitudes (900~700 hPa) should not be less than 0.2 g·kg^-1.The greater the updraft, the greater the water content of solid and liquid, and the greater the precipitation.The wind speed at 10 m observed by the automatic meteorological station drops to 0 m·s^-1 or there are no measured results.The time of freezing rain can be roughly estimated, and the results of manual observation can be effectively corrected or supplemented.

The error correction of the 10 m wind speed forecasted by the WRF model at 19 meteorological stations in Anhui province from April to December 2020 was carried out using the improved similar Kalman filter method.The results show that the average deviation of wind speed forecast is reduced from 1.35 m·s^-1 to 0.08 m·s^-1, which is about an elimination of the systematic error of the model.The root-mean-square error (RMSE) decreases from 1.77 m·s^-1 to 0.81 m·s^-1.When the average wind speed is more than 3 m·s^-1, the root-mean-square error of the wind speed forecast is reduced from 2.01 m·s^-1 to 1.19 m·s^-1, indicating that this method can not only effectively reduce the systematic error of the model, but also greatly reduce the random error of the model.The similar Kalman filter can correct the error of the physical process model which cannot be simulated accurately, improve the forecast accuracy of the model when the weather system changes dramatically and is suitable for the continuous forecast of meteorological elements for 24~72 hours.

Using the European center reanalysis data (ERA5), China Meteorological Administration mesoscale numerical prediction model (CMA-MESO) products and automatic station data, the convection parameters such as dynamic conditions, unstable stratification, water vapor conditions, and characteristic layer height were calculated, and the statistical characteristics of the quantile value of convection parameters for thunderstorm gale and short-term heavy rainfall, as well as the deviation characteristics of the climate value were counted.The results show that short-term heavy rainfall is more likely to occur in the middle and lower atmosphere which is nearly saturated in the humid and hot environment with high water vapor content.The uplift of the middle and lower layers triggers precipitation and enhances precipitation efficiency.Water vapor condition is the key to short-term heavy rainfall.Thunderstorm gale is easy to happen in the environment with large temperature drop rate, and especially for the conditions with dryness in the middle layer, wetness in the low layer, large vertical wind shear and large Cape.By using the relative deviation fuzzy matrix evaluation method, the classification forecast of two types of severe convective weather in Heilongjiang province is tested.The results show that the method can effectively predict the area and time in which severe convection is most likely to happen and has a good forecast effect and the reasonable rate of missing and falsity.The forecast BIAS is 0.7 for short-term heavy rainfall and 1.04 for thunderstorm gale.

In this study, three high resolution reanalysis precipitation datasets (CMA-RA, CMFD, and ERA5) were compared with annual, seasonal and monthly precipitation and the amount of Meiyu precipitation observed at weather station in Anhui province from 1979 to 2018 and their applicability were evaluated The results show that all three datasets can well describe the variation characteristics of seasonal precipitation in Anhui province, with most rainfall concentrated in summer, more rainfall in spring than in autumn, and least in winter.However, all the monthly precipitation is overestimated, with the maximum mean absolute percentage error (MAPE) found in December.In addition, precipitation in mountainous areas is significantly overestimated by the three datasets.Comparatively, the CMA-RA dataset agrees better with the observations than the other two datasets in describing the spatio-temporal distribution characteristics of the annual, summer and winter rainfall across the whole province and the Meiyu precipitation in regions along and south to Yangtze River.In general, the CMFD well captures the interannual variations of the precipitation on all timescales, but is rather poor in depicting their spatial distributions.However, it performs slightly better in describing the spatial-temporal characteristics of precipitation in spring and autumn and the interannual variation characteristics of Meiyu precipitation.As for the ERA5, the precipitation in Anhui province is systematically overestimated, with the bias being the largest among the three datasets, and it performs slightly better than other datasets in the precipitation in January.

This paper statistically analyzed spatial-temporal distribution features of triggered local convective weather under the control of the Western Pacific subtropical high (WPSH) in midsummer in the Yangtze Delta region from 2015 to 2019.The results show that the daily variation of the triggered local convection indicate a unimodal shape from noon to evening, and the high-frequency areas are located at the vicinity of large cities, the hilly areas, as well as the junction of water and land, but the convection rarely occurs over the rivers and lakes, canyons or basins.It is also found that high-frequency regions of the Southward type are mostly located in mountainous and coastal areas in the low-level westerly wind field, which is in favor to the surface heating.The high-frequency regions of the Northward type are located in inland, due to the low-level easterly wind weakening the surface heating and reducing the convection occurrence in the coastal area.The statistical analysis of the given convective parameters of two subtropical high types shows that the convective available potential energy (CAPE) value plays an obvious reference role in judging whether local convection is triggered.For coastal areas, the CAPE value of the Southward type is slightly higher than that of the Northward type, while it is opposite in inland areas.The lifting index (LI) and lifting condensation level (LCL) values of the Southward type are lower than that of the Northward type, reflecting intense stratification instability of the Southward type.

Using the data from automatic weather station, lightning locator, FengYun-2G satellite and ERA5 reanalysis data in Liaoning province from 2017 to 2021, the spatial-temporal distribution and environmental conditions of thunderstorm strong wind under different weather systems were analyzed.The results show that the cold vortex system is the most important weather system leading to thunderstorm strong winds in Liaoning province, accounting for 62% of all strong winds.Cold vortex thunderstorm strong winds are mainly concentrated in the central plain of Liaoning province, where cold vortex-induced thunderstorm strong winds account for 90% of all thunderstorm strong winds.Thunderstorm strong winds occur intensively from May to September, most frequently occurring from afternoon to evening.The environmental conditions of different seasonal thunderstorm strong winds vary with seasons.In spring and autumn, the thunderstorm strong winds correspond to stronger temperature differences at various levels, vertical wind shear and storm-bearing wind speeds.In the summer, the thunderstorm strong winds occur corresponding to abundant water vapor conditions, stronger near-surface evaporation, mid-level dry air intrusion and significant thermal convection feature.Compared with non-cold vortex systems, under the cold vortex background, the low-level to 500 hPa temperature difference and storm-bearing wind speeds are greater, the mid-low layer is drier, and the vertical wind shear and convective available potential energy are stronger.Among the four quadrants of the cold vortex, thunderstorm strong winds often occur in the southeast and southwest quadrants, accounting for 77.9% and 18.5% respectively.There are differences in the environmental conditions corresponding to the thunderstorm strong winds in the two quadrants.The southeast quadrant has more abundant water vapor conditions and stronger instability.The mid-low layer is drier in the southwest quadrant, and the temperature difference between surface and 500 hPa and the downdraft convective available potential energy are greater.

Using cluster analysis, potential source analysis (PSCF, CWT) and obliquely rotated T-mode principal component analysis (PCT) method, the transmission channels and potential source areas were studied, and the prone circulation patterns was revealed for different stages (formation, maintenance and dissipation) of heavy air pollution in Huaibei in winter half year, based on the data of the monthly variation of six types of air pollutants from 2015 to 2018.The results show that monthly average mass concentrations of the main air pollutants, i.e.PM_2.5, PM₁₀, CO, NO₂ and SO₂ are higher in winter half year.For heavy pollution in Huaibei, short distance trajectory is dominant, and the important potential sources include the southeast of west of Shandong province and north central of Jiangsu province.There are similarities and difference during the different stages (formation, maintenance, and dissipation) of heavy air pollution.The common point is that there are uniform pressure field with different proportions in the three stages.The differences among the three stages are mainly expressed in circulation type, pressure value, pressure gradient, wind direction and wind speed.In the formation stage of persistent heavy pollution, circulation patterns are divided into five types and more than 90% of them are related to uniform pressure field.In the maintenance stage, six circulation patterns are related to uniform pressure field, but the control range of uniform pressure field is significantly larger than that in the formation stage.The pressure value was close to the formation stage and the wind speed was mostly static.In the dissipation stage of persistent heavy pollution, Huaibei city was mostly in the front of cold high pressure, with the uniform pressure field just accounting for 15%.The pressure gradient, sea-level pressure, and wind speed increased significantly.

Using encrypted automatic stations, reanalysis data and satellite fog retrieval data, the spatial and temporal distribution characteristics of sea and land fog on the west coast of Bohai Sea from 2015 to 2020 were analyzed, and a typical case was selected to carry out sensitivity numerical experiments.The results show that the number of land fog is greater than that of the sea fog in autumn and winter on the west coast of Bohai Sea.The opposite results are true in spring and summer.Although the amount of sea fog and land fog in winter are the largest and the variation trend is the same, the distribution of land and sea fog still have significant local characteristics.The probability of fog bounded by the coastline on the west coast of Bohai Sea is more than 53%, most of them occur at the night, and the duration time is less than 12 h.The case analysis shows that the sea and land wind circulation transport the warm and humid air from the sea to the low altitude of the inland, the humidity increases rapidly to near saturation, and the weak cold air intrusion and the cooling due to the long wave radiation at night work together, leading to the formation of the inland fog.However, the sea fog is difficult to form due to high sea surface temperature (SST) and slow cooling of ocean atmosphere.The results of the SST sensitivity test also show that when the SST increases by 5%, the fog area that slightly enters the offshore area in the control experiment basically rewinds to the west of the coastline, and when the SST decreases by 5%, the land fog spreads to the whole Bohai Sea and maintains.High SST is the key factor that the land fog does not expand into the sea.

Using 199 meteorological stations in Northeast China from 1981 to 2020, the temperature, precipitation, atmospheric circulation field and sea surface temperature (SST) field from 1981 to 2010 were compared with those from 1991 to 2020, and the influence of the change of climate mean values on climate evaluation, climate change and climate prediction was analyzed.The results show that there are clear changes in temperature, precipitation, atmospheric circulation and SST field in northeast China under new and old climate states, and the changes of circulation and SST field correspond to the changes of temperature and precipitation.Under the new climate state, the precipitation in spring, autumn and winter increases, while in summer it decreases.The average temperature increases.The atmospheric circulation field, the high latitude blocking high pressure and the West Pacific Subtropical high pressure increase in winter, and the East Asia trough weakens, corresponding to the increase in winter temperature.In summer, the Okhotsk Sea blocking high pressure and the Western Pacific Subtropical High pressure weaken, which corresponds to the decrease of summer precipitation in Northeast China.As for the SST field, the global SST is generally increasing, while those in the middle and high latitudes in the southern Hemisphere are decreasling.The summer SST difference in the equatorial Pacific region is positive in the west and negative in the east, which is consistent with the rise of temperature in summer in the northeast.

Using the daily precipitation data at 204 meteorological stations in Northeast China, monthly NCEP/NCAR reanalysis data, and Niño3.4 index provided by the National Climate Prediction Center (CPC) of the United States, the variation characteristics of summer precipitation as well as its relationship with the previous ENSO in Northeast China were analyzed by correlation analysis, wavelet analysis, M-K mutation, composite analysis and other methods.The results show that the annual precipitation in summer shows a decreasing trend at the rate of 2.39 mm/10 a in Northeast China.It decreases in southwest and increases in northeast.The highest decreasing rate occurs in Liaoning region.Summer precipitation has significant oscillation periods of 2~3 a, 6~7 a and 8~9 a.1961 to 1982 and 1998 to 2019 are relatively dry periods, 1983 to 1997 are relatively wet periods, and the abrupt change years are 1983 and 1998.Summer precipitation in Northeast China is closely related to the summer ENSO of the previous year, and there is a significant positive correlation between the two from 1997 to 2019.From 1997 to 2019, SST anomaly changes in Niño3.4 region in the previous summer caused water vapor transport anomaly and local vertical movement changes, which has a significant impact on summer precipitation in Northeast China.Niño3.4 index can be used as a predictor of summer precipitation in Northeast China in the following year.

According to the weather type, PM_2.5 pollution days in the Beijing-Tianjin-Hebei region from 2015 to 2018 were divided into uniform pressure type, rear part of a high pressure type and northwest high-pressure type.WRF-CAMx and WRF-CMAQ models were used to analyze PM_2.5 sources and simulate emission reduction in Beijing and Shijiazhuang.The results show that compared with the pressure equalization type and the northwest high-pressure type, PM_2.5 pollution under the rear part of a high pressure type appears least in Beijing and Shijiazhuang, but the transport of external pollution is more obvious there.On PM_2.5 pollution days under three weather types, the emission contribution of local pollution sources in Shijiazhuang is higher than that in Beijing, which is one of the main reasons why the PM_2.5 concentration in Shijiazhuang is higher than that in Beijing.Affected by different weather types, the transport and source characteristics of PM_2.5 pollutants are different in Beijing and Shijiazhuang.According to the transport characteristics of PM_2.5 pollutants in Shijiazhuang, the emission reduction optimization scheme can reach 83.6%~91.3% of the joint emission reduction of the central and southern cities of Beijing-Tianjin-Hebei region, which is reasonable and efficient.

Taken spring maize in Northeast China as the research object, based on the experimental data of Jinzhou Agroecological Observation Station from 2018 to 2021, parameters of CERES-Maize crop growth model were calibrated and verified.The effects of irrigation and nitrogen management on yield, water use efficiency and economic income of spring maize in different precipitation years (normal, dry and wet years) were simulated to find the optimal water and nitrogen management scheme for spring maize in Northeast China under different precipitation years and production targets.The results show that the normalized root mean square error of CERES-Maize model in simulating the growth period and yield of spring maize is less than 10%, which can meet the requirements of the simulation accuracy of the growth process and yield of maize.The highest yields of spring maize under different irrigation and nitrogen management in different precipitation years are 1.08×10⁴, 1.16×10⁴, 1.13×10⁴ kg·ha^-1 respectively, which is highest in dry year and lowest in normal year.The highest yields are obtained by irrigation quantity of 185, 205, 175 mm, and nitrogen application amount of 75, 175, 125 kg·ha^-1, respectively.The maximum water use efficiency of spring maize under different water and nitrogen management in different precipitation years is the highest in wet year and the lowest in dry year.The highest water use efficiency is obtained by irrigation quantity of 175, 195, 175 mm, and nitrogen application amount of 75, 175, 125 kg·ha^-1, respectively.The maximum economic benefits of spring maize under different water and nitrogen management in different precipitation years are the largest in dry year and the smallest in normal year.The maximum economic benefits are obtained by irrigation quantity of 175, 205, 175 mm, and nitrogen application amount of 75, 175, 125 kg·ha^-1, respectively.

Using daily precipitation and daily average temperature from 52 meteorological stations in Liaoning province from 1961 to 2020, the standard precipitation index (SPI) and the standard precipitation evaporation index (SPEI) were calculated at soybean growing season scale and at each developmental stage scale.Mann-Kendall test was used to examine the spatial and temporal variation characteristics of drought and drought frequency of soybean at different time scales.The study showed that both SPI and SPEI indices showed aridity trends during soybean growing season, with SPEI monitoring a slightly more severe drought than SPI.At all soybean developmental stages, SPEI and SPI indices showed wetting trends at the sowing-emergence stage, and aridity trends at the emergence-flowering, flowering-podding, and podding-maturity stages, with significant aridity trends at the flowering-podding stage.Spatially, the SPEI index monitored a larger range of drought and drought prone areas than those monitored by the SPI, and the degree of drought was stronger, with the most severe drought happening in the west of Liaoning.Therefore, the SPEI index is more applicable compared to the SPI index when evaluating drought of the soybean at growing season and each developmental stage.

Lightning distribution characteristics in Liaoning province such as intensity, density and lightning current amplitude, and their relationship with the altitude were analyzed using the lightning location data in Liaoning province from 2010 to 2021, combined with space natural breakpoint classification method.The results show that the amplitude distribution of lightning current in Liaoning province is concentrated in summer, and the proportion of positive lightning distribution is the lowest.The amplitude of lightning current in winter is significantly higher at daytime than that at night.Except for August, the lightning frequency in other months is usually higher at daytime than that at night.In western Liaoning, the positive and negative lightning density is generally large, and the positive and negative lightning density decreases with the increase of altitude, and the negative lightning density decreases faster than positive lightning.The amplitude of positive lightning is the highest in western Liaoning province, and the amplitude of positive lightning decreases with the increase of altitude, while the amplitude of negative lightning generally increases with the increase of altitude.The high value area of positive lightning intensity is mainly distributed in the middle and west of Liaoning province, and the high value area of negative lightning intensity is mainly concentrated on both sides of Liaohe River Basin, with the strongest distribution in the east of Liaoning province.On the profile of 41°N, the positive flash density in western Liaoning province is much higher than that in Eastern Liaoning province, and the negative flash density in hilly area of western Liaoning province is significantly lower than that in Liaohe plain in central Liaoning province.In the hilly region of the western Liaoning province, the amplitude change of negative lightning is consistent with the change of altitude.The intensity of positive lightning is the highest in western Liaoning province, the lowest in Eastern Liaoning province, and the intensity of negative lightning is the strongest in the windward slope area in the east of Liaohe plain.

To understand the evolution characteristics and cloud physical mechanism of lightning for hail and heavy rain clouds, using VLF/LF three-dimensional lightning monitoring data, combined with S-band dual polarization radar and ground observation data, statistical and comparative analysis methods were used to analyze the characteristics of individual lightning for 31 hails and 32 short-term heavy rains.The characteristics of single lightning, dual polarization radar echo and the correlation between lightning and convective development height were analyzed and compared in detail.The results show that more than two-thirds of the hail samples with lightning frequency are more than 50 times/6 min before hails fall.The ratio of positive ground lightning and positive cloud lightning of a hail monomer is high, while that of mostly heavy precipitation monomer is very low.The peak value of lightning frequency of 80% individual hail is 2~25 minutes ahead of the time of hails fall.The total lightning frequency of most individual hail rises sharply before hails fall, and the increasing rate of frequency is more than 4 times/min.The peak value of single lightning frequency is 2~35 minutes ahead of the precipitation peak value in more than half of the heavy rainfall, and the lightning frequency increases before the precipitation peak value in most single precipitation, and the increasing rate of lightning frequency is mostly below 4 times/min.The cloud flash frequency of the two types of monomers in the mature stage is the highest among the three stages.The cloud flash in the three stages is concentrated in the layer of 2~6 km height.In the hail monomer, the lower layer below the melting layer is composed of hail and heavy rain particles, and the upper layer is composed of hail and graupel.In the heavy rain monomer, the lower layer below the melting layer is composed of heavy rain particles, and the upper layer is composed of ice crystals and supercooled water droplets.The higher the convective height is, the more obvious the ice phase process is and the stronger the lightning activity is.In this study, lightning data and dual polarization radar parameters were combined to provide a reference for the identification and early warning of different types of severe convective weather and lightning.

Based on the CMPAS (CMA Multi-source merged Precipitation Analysis System) algorithm developed by National Meteorological Information Center(NMIC), using the ground observational precipitation data and radar quantitative precipitation estimation data, Sichuan provincial precipitation products of 0.01°/1 h (SC) during April to September, 2021, were produced.Taking the hourly precipitation observation data as true value, SC and real-time precipitation products of 0.01°/1 h (RT) developed by NMIC were evaluated.The evaluation was divided into two categories according to whether the observation data is modified or not.On this basis, using the machine learning support vector regression algorithm (SVR), the improvement method of RT was developed.The results show that, in the case of observation data is not modified, RT is prone to overestimation and SC is prone to underestimation.These two kinds of products have similar performance in the basin area.RT has a better performance in Panzhihua, Liangshan, Ganzi and Aba region.In the case in which the observation data are modified, no matter the whole province or each region within the province, SC is clearly better than RT.RT is significantly larger than the observation value.This implies that the wrong observational data before modification have a significant impact on the performance of RT.The mean absolute error of the improved products is reduced by 29.9% and the root mean square error is reduced by 41.1% and the correlation coefficient is increased by 4.94% compared with that before the improvement.

Using the observational data of daily temperature from 76 national meteorological stations in Hebei province from March to May during 1991-2020, the spatial and temporal characteristics of late spring coldness were analyzed using the methods of Mann-Kendall test, GIS spatial interpolation and linear regression.The results showed that the accuracy of "Meteorological index of late spring coldness" in screening the process of late spring coldness was 63.6%, with satisfied applicability.The number of late spring coldness occurrence was 2094 from 1991 to 2020 in Hebei province, with the mild occurring the most, followed by moderate, and the severe occurring the least.In terms of the spatial distribution, the number of the mild late spring coldness increased from south to north, the moderate-severe increased from the middle to the north and the south, with the most distributed in the northern plateaus and mountains.The inter-annual variation of late spring coldness showed a downward trend, with the most obvious decline happening in central Hebei province.In terms of total duration days, the mild decreased from northeast to southwest, and moderate-severe decreased from the northern plateau, Yanshan mountain area, Taihang mountains to the plains.The inter-annual variation of total duration days of late spring coldness in southern and central Hebei province showed a downward trend.The daily minimum temperature decreased in a stepwise manner from south to north, with minimum temperature in the north of Hebei province far below the south and central of Hebei province.The degree of daily temperature reduction decreased in a stepwise manner from south to north in March and May.In central Hebei province, the number of moderate-severe late spring coldness was large from April 9 to 12, while in the north of Hebei province, the total number of late spring coldness and the number of moderate-severe in early May were relatively high.