Meteorological station surrounding snow parameter observation and analysis method

Abstract

The present application relates to satellite remote sensing technical field, specifically, a kind of meteorological station periphery snow parameter observation and analysis method, including snow profile observation, snow profile observation and unmanned aerial vehicle observation, the summary analysis of observation result, snow spatial distribution heterogeneity analysis, compared with prior art, by snow profile observation, snow profile observation and unmanned aerial vehicle observation, obtain the spatial distribution characteristics of parameter such as snow depth, snow water equivalent, snow grain size, section snow density, snow water content, stratification and the digital observation data of larger area in observation area, then from snow depth, snow water equivalent and snow density three aspects, the snow spatial distribution heterogeneity and spatial representation of the region are analyzed, realize the positive driving effect to snow parameter inversion algorithm, snow product verification, snow depth, snow water equivalent service etc.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of satellite remote sensing, in particular to a meteorological station surrounding snow parameter observation and analysis method. BACKGROUND

[0002] Snow is an important link in global energy-balance, after large-scale snowfall, how to obtain reasonable and reliable snow-related parameters by various means has been concerned by relevant departments.

[0003] Meteorological station snow monitoring data can be obtained stably for a long time, through the observation of meteorological stations, the spatial characteristics of snow distribution after snowfall can be obtained to a certain extent, but the distribution of snow is closely related to meteorological elements, topographic elements, surface vegetation, water distribution and other factors, which causes the snow to be unevenly distributed in space.

[0004] Satellite remote sensing snow monitoring can quickly obtain snow distribution characteristics of a large range and multiple scales, but the universality of satellite remote sensing snow parameter inversion algorithm still has some problems, which mainly manifest in that the snow characteristics are different for different regional underlying surface characteristics and climate characteristics, so there will be various problems in the process of popularizing the snow parameter inversion algorithm established for a region to other regions; and satellite remote sensing monitors surface information, and most of the ground observation data used to establish satellite remote sensing snow inversion algorithm are point observation results, so whether the point observation results are representative in the corresponding satellite remote sensing observation pixels is also one of the important factors affecting the establishment of satellite remote sensing snow inversion algorithm.

[0005] The Chinese regional algorithm in the satellite microwave imager snow depth and snow water equivalent algorithm is developed by using ground station observation data combined with satellite brightness temperature observation data, through comparison with the station observation data, it is found that the verification results of the algorithm at some stations have relatively significant errors, therefore, how to further improve the algorithm accuracy becomes a problem that must be considered. From the source of the algorithm development, if the accuracy of the input parameters in the algorithm development process can be improved, it will help to improve the algorithm, for the Chinese regional algorithm, in addition to satellite data, the main input parameter is the snow depth monitoring data of meteorological stations, therefore, the spatial representativeness of each meteorological station snow depth monitoring becomes a problem that must be considered, in the next step of algorithm improvement process, the data of stations with good spatial representativeness will be retained, and the data of stations with poor spatial representativeness will be considered to reduce the weight or discarded according to the spatial representativeness.

[0006] Therefore, the application provides a meteorological station surrounding snow parameter observation and analysis method, which is aimed at the snow depth and snow water equivalent of the surrounding of the key meteorological station, and observes other snow elements such as snow particle size, snow density and snow layering, so as to obtain the spatial representative condition of the snow parameter observation result of the related meteorological station. SUMMARY

[0007] The application aims at overcoming the defects of the prior art, and provides a meteorological station surrounding snow parameter observation and analysis method, which is aimed at the snow depth and snow water equivalent of the surrounding of the key meteorological station, and observes other snow elements such as snow particle size, snow density and snow layering, so as to obtain the spatial representative condition of the snow parameter observation result of the related meteorological station.

[0008] In order to achieve the above object, the application provides a meteorological station surrounding snow parameter observation and analysis method, which comprises snow profile observation, snow profile observation and unmanned aerial vehicle observation, observation result summary analysis and snow spatial distribution heterogeneity analysis, and specifically comprises the following steps.

[0009] A, snow profile observation: the spatial distribution characteristics of the snow depth, snow water equivalent and snow particle size are obtained, including snow depth profile observation and snow depth, snow water equivalent and snow particle size profile observation, the snow depth profile observation is used to obtain the spatial distribution condition of the snow depth to the greatest extent, and the snow depth, snow water equivalent and snow particle size profile observation is used to extend the result to the range of 25km based on the snow depth spatial distribution result, in cooperation with the observation of the snow water equivalent and snow particle size.

[0010] B, snow profile observation: including snow density, snow water content and layering, except that the natural layering needs to be increased, the observation of the snow characteristics is performed from the top layer of the snow every 5cm until the ground.

[0011] C, unmanned aerial vehicle observation: the unmanned aerial vehicle is used to perform aerial observation on the snow-covered ground, the obtained all ground shooting results of each area flight are spliced to obtain the overall observation result, the DEM digital elevation matrix, DOM digital orthographic image and DSM digital surface model result are obtained, and the scale conversion from point to plane is realized.

[0012] D, observation result summary analysis: including meteorological station surrounding snow spatial distribution heterogeneity analysis and meteorological station spatial representative analysis.

[0013] The spatial heterogeneity of snow around the meteorological station is analyzed by the snow depth, snow water equivalent and snow density. The spatial distribution standard deviation of each snow parameter is quantitatively estimated by the ratio of the spatial distribution standard deviation to the observation mean value.

[0014] The spatial heterogeneity index of snow depth: SD Het = SD Stdev / SD Avg

[0015] The spatial heterogeneity index of snow water equivalent: SWE Het = SWE Stdev / SWE Avg

[0016] The spatial heterogeneity index of snow density: Den Het = Den Stdev / Den Avg

[0017] Het is the spatial heterogeneity index of snow;

[0018] According to the spatial heterogeneity index:

[0019] For the inversion of snow parameters, the spatial representativeness of the meteorological station needs to be evaluated for the specific snow parameter to be inverted. The spatial representativeness of the observed snow depth of the meteorological station is included in the algorithm for directly inverting the snow depth and calculating the snow water equivalent using the snow density lookup table. The quality identification weight of the station snow depth observation result is used to evaluate and weight the station data entering the algorithm;

[0020] For the snow water equivalent algorithm using station snow density data for assimilation, the spatial representativeness of the observed snow density of the meteorological station is included in the algorithm, and the quality identification weight of the station snow density observation result is used to evaluate and weight the station observation data entering the algorithm;

[0021] The weight is obtained through ground observation test or through cooperation with meteorological station through daily spatial encryption observation work;

[0022] Meteorological station spatial representativeness analysis: the reliability of the meteorological station site selection in the observation of snow parameters is evaluated. The observation results of the meteorological station site accurately describe the large-scale distribution average value of the snow parameters, including three groups of indexes for evaluating the snow depth observation data of the meteorological station:

[0023] The absolute deviation of snow depth of the meteorological station: SD AbDe = SD Station - SD Avg

[0024] The relative deviation of snow depth of the meteorological station: SDReDe = (SD Station - SD Avg ) / SD Avg

[0025] Snow depth weather station location index:

[0026] According to three groups of indexes, the snow depth observation results are calculated, according to the calculation results, in the actual station data using process, the difference of the natural snow state distribution around the station and the difference of the spatial representativeness of the station are distinguished;

[0027] The difference of the natural snow state distribution around the station mainly considers the mixed pixel problem in the development of the snow algorithm;

[0028] The difference of the spatial representativeness of the station mainly considers the precision problem in the verification of the snow depth;

[0029] E, snow spatial distribution heterogeneity analysis:

[0030] Including the influence of wind speed and direction on snow redistribution and the influence on the spatial representativeness of the snow parameter observation results of the weather station;

[0031] From the observation of the ground snow profile, the influence of wind speed and direction on snow is analyzed, through the snow density of each layer of the profile being higher than the average value of the snow density around the station, and most of the layered snow does not show the characteristics that the snow density increases with the increase of depth, it is shown that the abnormal snow density of the profile snow layer is not caused by the gravity of the snow itself, and the snow water content does not show the characteristics that the snow water content gradually increases with the increase of snow depth, and the influence of wind blowing snow on snow redistribution is obtained by comparing the profile layering characteristics with other snow layering characteristics;

[0032] Through the relationship between the maximum and average values of the wind speed of each station and the snow parameters and the relationship between the station wind direction and the snow parameters, the influence of wind speed and direction on snow redistribution is obtained, and then the spatial representativeness of the snow parameter observation results of the weather station is affected.

[0033] Compared with the prior art, through snow profile observation, snow profile observation and unmanned aerial vehicle observation, the spatial distribution characteristics of snow depth, snow water equivalent, snow particle size and other parameters, snow density, snow water content, layering and digital observation data of a larger observation area are obtained, and then the snow spatial distribution heterogeneity and spatial representativeness of the region are analyzed from three aspects of snow depth, snow water equivalent and snow density, which realizes the positive promoting effect on the snow parameter inversion algorithm, snow product verification, snow depth, snow water equivalent service and other work. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1A space diagram of snow depth profile on the UAV aerial photography result (Habahe observation point) for the embodiment of the present application is shown in the figure.

[0035] Figure 2 A space diagram of snow spatial distribution heterogeneity around the weather station for the embodiment of the present application is shown in the figure.

[0036] Figure 3 A space diagram of snow depth spatial representation of the weather station in northern Xinjiang for the embodiment of the present application is shown in the figure.

[0037] Figure 4 A diagram of the maximum wind speed of each weather station in northern Xinjiang in December, January and February from 1951 to 2015 for the embodiment of the present application is shown in the figure.

[0038] Figure 5 A diagram of the average wind speed of each weather station in northern Xinjiang in December, January and February from 1951 to 2015 for the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0039] The present application will be further described in conjunction with the accompanying drawings.

[0040] Reference Figures 1-5 , the present application provides a kind of meteorological station surrounding snow parameter observation and analysis method, including snow profile observation, snow profile observation and unmanned aerial vehicle observation, the summary analysis of observation result, snow spatial distribution heterogeneity analysis, specifically:

[0041] A, snow profile observation: obtain the spatial distribution characteristics of snow depth, snow water equivalent, snow particle size and other parameters, including snow depth profile observation and snow depth, snow water equivalent, snow particle size profile observation, snow depth profile observation is to obtain the spatial distribution of snow depth to the greatest extent, snow depth, snow water equivalent, snow particle size profile observation is based on the result of snow depth spatial distribution, with the observation of snow water equivalent and snow particle size, the result is extended to 25km scale range;

[0042] B, snow profile observation: including profile snow density, snow water content, stratification, except for obvious natural stratification needs to increase observation stratification, the rest from the top layer of snow, every 5cm carries out once snow property observation, until the ground;

[0043] C, unmanned aerial vehicle observation: using unmanned aerial vehicle to carry out aerial photography observation of snow-covered ground, splicing all the ground shooting results obtained by each area voyage to obtain the overall observation result, obtaining DEM digital elevation matrix, DOM digital orthographic image and DSM digital surface model result, realizing the scale conversion from point to plane;

[0044] D, Summary analysis of observation results: including spatial heterogeneity analysis of snow around weather stations and spatial representativeness analysis of weather stations:

[0045] The spatial heterogeneity analysis of snow around weather stations is quantitatively estimated by the ratio of the spatial distribution standard deviation of each snow parameter to the average value of the observation through three snow parameters of snow depth, snow water equivalent and snow density:

[0046] Snow depth spatial heterogeneity index: SD Het = SD Stdev / SD Avg

[0047] Snow water equivalent spatial heterogeneity index: SWE Het = SWE Stdev / SWE Avg

[0048] Snow density spatial heterogeneity index: Den Het = Den Stdev / Den Avg

[0049] Het is the spatial heterogeneity index of snow;

[0050] According to the results of the spatial heterogeneity index:

[0051] For snow parameter inversion, the spatial representativeness of the snow station needs to be evaluated for the specific snow parameter required for inversion. Directly invert the snow depth, use the snow density lookup table to calculate the snow water equivalent algorithm, and include the spatial representativeness of the weather station observation snow depth into the algorithm. The quality identification weight of the station snow depth observation result is used to evaluate and weight the station data entering the algorithm;

[0052] For snow water equivalent algorithm using station snow density data for assimilation, the spatial representativeness of the weather station observation snow density is included in the algorithm, and the quality identification weight of the station snow density observation result is used to evaluate and weight the station observation data entering the algorithm;

[0053] The weight is obtained through ground observation test or through cooperation with weather station through daily spatial encryption observation work;

[0054] Spatial representativeness analysis of weather stations: evaluate the reliability of weather station site selection in snow parameter observation, and the observation results of the location of weather station can accurately describe the average value of large-scale distribution of snow parameters, including three groups of indexes to evaluate the snow depth observation data of weather station:

[0055] Snow depth absolute deviation of weather station: SD AbDe = SDStation -SD Avg

[0056] Relative deviation of snow depth observation of meteorological station: SD ReDe = (SD Station - SD Avg ) / SD Avg

[0057] Position index of snow depth observation of meteorological station:

[0058] According to the three groups of indexes, the snow depth observation results are calculated, and according to the calculation results, in the actual station data using process, the difference of natural snow state distribution around the station and the difference of spatial representativeness of the station are distinguished;

[0059] The difference of natural snow state distribution around the station mainly considers the mixed pixel problem in the development of snow algorithm;

[0060] The difference of spatial representativeness of the station mainly considers the precision problem in the verification of snow depth;

[0061] E. Analysis of snow spatial distribution heterogeneity:

[0062] Including the influence of wind speed and wind direction on snow redistribution and the influence on the spatial representativeness of snow parameter observation results of meteorological station;

[0063] From the observation of the ground snow profile, the influence of wind speed and wind direction on snow is analyzed, through the snow density of each layer of the profile being higher than the average value of the snow density around the station, and most of the layered snow does not show the characteristics that the snow density increases with the increase of depth, it is shown that the abnormal snow density of the profile snow layer is not caused by the gravity of the snow itself, and the snow water content does not show the characteristics that the snow water content gradually increases with the increase of snow depth, and through the comparison of the profile layering characteristics with other snow layering characteristics, the influence of wind blowing snow on snow redistribution is obtained;

[0064] Through the relationship between the maximum and average values of wind speed of each station and the snow parameters, and the relationship between the fixedness of wind direction of each station and the snow parameters, the influence of wind speed and wind direction on snow redistribution is obtained, and then the spatial representativeness of the snow parameter observation results of meteorological station is affected.

[0065] Embodiment:

[0066] The application is aimed at the main weather stations in northern Xinjiang, including Beitashan, Qinghe, Fuyun, Alatai, Jimunai, Habax, Tuoli, Yumin, and Tacheng. The above-mentioned stations cover the main snow-covered areas north of the Tianshan Mountains in northern Xinjiang. Compared with the Ili River Valley area, the weather conditions for snowfall and snowmelt are relatively consistent, and the redistribution of snow mainly depends on the ground conditions, which can better analyze the impact of ground conditions on snow redistribution.

[0067] This test basically covers the main weather stations in northern Xinjiang except the Ili River Valley. For each weather station, routine observations such as snow depth profile (SD) observation, snow water equivalent and snow particle size profile (SP) observation, and snow profile observation are carried out. In addition, unmanned aerial vehicle observation is also carried out in the observation area. The two weather stations of Beitashan and Qinghe involved in the test mainly carried out detailed snow depth profile observation and unmanned aerial vehicle observation. From Fuyun, comprehensive routine observations including snow depth, snow water equivalent, and snow particle size are carried out in the surrounding area of the weather station.

[0068] The main content of this experiment includes three aspects

[0069] (1) Snow profile observation

[0070] (2) Snow profile observation

[0071] (3) Unmanned aerial vehicle observation

[0072] Around the weather station, set up profile observation area in 3-4 directions, 2 observation areas in each direction, 5km apart, covering an area of 20-25km. Carry out snow depth profile observation in the set observation area. The profile needs to be perpendicular to the road direction, and different profile observation areas should include different terrains, rivers, and underlying surfaces. Every 20 meters (30 steps) in the profile, take 3 sets of snow depth data, take the average as the snow depth result of this point, a total of 20-50 sets of snow depth observation, profile length 400-1000 meters, specific conditions according to the observer and snow conditions. In the actual observation process, one person holds a ruler to observe the snow depth, and the other person holds a notebook and GPS to record the average snow depth read by the observer and the GPS point number of this point. In the data summary process, the actual latitude and longitude data are derived according to the GPS number and sorted out.

[0073] The actual observation results are displayed on the information of the unmanned aerial vehicle flight results Figure 2 ), the results of snow depth profile observation are reliable, the profile direction is stable, and the spacing between observation points also meets the requirements.

[0074] Centered on the meteorological station, profile observation zones are set up in 3-4 directions, with two observation zones 5km apart in each direction, covering an overall area of ​​20-25km. These zones are alternated with snow depth profiles to ensure 12-16 sets of profile observation data around each meteorological station. Within the designated observation zones, snow depth, snow water equivalent, and snow grain size profiles are observed. The profiles must be perpendicular to the road direction, and different profile observation zones should incorporate different terrains, rivers, and underlying surfaces as much as possible. Observations are taken every 20 meters (30 steps) within the profile, with 3 sets of snow depth, snow water equivalent, and snow grain size data collected each time. The average is taken as the observation result for that point, for a total of 5 sets of snow depth observations. The profile length is 100 meters. In actual observation, snow depth, snow water equivalent, and snow grain size profile observations require two people working together. One person is responsible for operating the snow tube to obtain snow depth and snow water equivalent, conducting three sets of observations at each point and reporting the results to the recorder. The other person is responsible for data recording and snow grain size observation.

[0075] This application employs two main methods for observing snow particle size: snow particle size estimation using a graduated microscope and snow specific surface area observation using the ICECube. In actual observations, a comparison revealed an uncertainty between the observed snow specific surface area and the snow particle size estimation results using a graduated microscope. Therefore, this application still uses the traditional microscope observation results as the final observed data for snow particle size.

[0076] Select 1-2 locations in each direction around the target weather station and conduct snow profile observations based on the snow cover conditions. After selecting the snow profile locations, first use a large shovel to dig a snow pit, ensuring the SnowFork can be inserted horizontally into the snow layer, while allowing the observer to maintain a suitable observation posture within the pit. Then, quickly clear the snow profile with a small snow shovel and begin observations immediately. All observations must be completed within 10 minutes to prevent changes in the snow cover characteristics of the snow profile due to environmental influences.

[0077] While digging the snow pit, the SnowFork operator needs to take out the instrument, connect the cables, and place the snow fork in the air beforehand to make it reach the same temperature as the air, so as to avoid the temperature difference between the snow fork and the snow layer causing observation errors.

[0078] The observation elements of the snow profile include snow density, snow moisture content, and stratification. In this application, except for obvious natural stratification that requires additional observation layers, the snow characteristics are observed every 5 cm from the top layer of snow until the ground surface.

[0079] This paper is the first attempt to use unmanned aerial vehicles for aerial observation of snow-covered ground. The model used is DJI-Wu, a small four-rotor unmanned aerial vehicle that can be operated by a single person. Under the conditions of airspace management in China, it can fly below 500 meters. Since there is no matching ground station software, the operation of the unmanned aerial vehicle in this test must rely on manual control by the operator. With stable height and flight direction, the operator needs to control the flight route and manually take pictures at fixed intervals using a remote control handle. The results of all ground shots obtained during each flight are then spliced together to obtain the overall observation results.

[0080] During the entire experiment, the unmanned aerial vehicle carried out 50 flights for observation, a total of more than 8000 original photos. In the later stage, professional unmanned aerial vehicle data splicing software was used to obtain 40 complete spliced orthographic images. This test shows that using unmanned aerial vehicles for observation can obtain a large area of digital observation data in a short time. After splicing, DEM (Digital Elevation Matrix), DOM (Digital Orthographic Image), and DSM (Digital Surface Model) results can be obtained. To some extent, it can achieve the scale conversion from point to plane. However, this test also exposes some problems in unmanned aerial vehicle observation. First, unmanned aerial vehicle observation still needs to be equipped with a relatively professional ground station software to plan the observation path and height. Second, the flight height in this test is relatively low, and the observation area is still small. After each observation and splicing, only 1 km 2 left and right,

[0081] If higher airspace is applied, better observation results can be obtained. Finally, the battery of the unmanned aerial vehicle is still the biggest obstacle to its operation in low-temperature conditions. In an environment of minus 20-30 degrees Celsius, each unmanned aerial vehicle battery can only work normally for about 10 minutes. In order to ensure the safety of the equipment, it is difficult to conduct larger-scale flight observations. In summary, unmanned aerial vehicle observation is a new trend of development, and it is worth further studying in future observation tests.

[0082] The spatial distribution heterogeneity of snow in the areas of the nine weather stations involved in this test was analyzed in terms of snow depth, snow water equivalent, and snow density. To quantitatively estimate the spatial heterogeneity of snow, we defined the snow spatial heterogeneity index Het, which was calculated using the ratio of the standard deviation of the spatial distribution of each parameter to the average value of the observation.

[0083] Snow depth spatial heterogeneity index: SD Het = SD Stdev / SD Avg

[0084] Snow water equivalent spatial heterogeneity index: SWE Het = SWEStdev / SWE Avg

[0085] Index of spatial heterogeneity of snow density: Den Het = Den Stdev / Den Avg

[0086] According to the above indexes, the observation contents of this test were calculated, and the calculation results are shown in Table 3. Figure 3 Wherein, the subscripts l and p correspond to the profile mean value and the original point respectively.

[0087] In terms of snow depth:

[0088] The station with the weakest spatial heterogeneity of snow depth obtained by profile mean value is Qinghe, followed by Fuyun, and then Tacheng. The station with the strongest spatial heterogeneity is Tuoli, followed by Jimunai, and then Hobq.

[0089] The station with the weakest spatial heterogeneity of snow depth obtained by original point observation results is Qinghe, followed by Fuyun, and then Tacheng. The station with the strongest spatial heterogeneity is Hobq, followed by Jimunai, and then Tuoli.

[0090] In terms of snow water equivalent:

[0091] The station with the weakest spatial heterogeneity of snow water equivalent obtained by profile mean value is Tacheng, followed by Qinghe, and then Beita Mountain. The station with the strongest spatial heterogeneity is Hobq, followed by Tuoli, and then Yumin.

[0092] The station with the weakest spatial heterogeneity of snow water equivalent obtained by original point observation results is Tacheng, followed by Qinghe, and then Fuyun. The station with the strongest spatial heterogeneity is Hobq and Jimunai, followed by Tuoli.

[0093] In terms of snow density:

[0094] The station with the weakest spatial heterogeneity of snow density obtained by profile mean value is Beita Mountain, followed by Tacheng, and then Qinghe. The station with the strongest spatial heterogeneity is Yumin, followed by Jimunai, and then Hobq.

[0095] The station with the weakest spatial heterogeneity of snow density obtained by original point observation results is Tacheng, followed by Qinghe, and then Beita Mountain. The station with the strongest spatial heterogeneity is Jimunai, followed by Hobq, and then Yumin.

[0096] The station with the weakest spatial heterogeneity of snow density obtained by original point observation results is Tacheng, followed by Qinghe, and then Beita Mountain. The station with the strongest spatial heterogeneity is Jimunai, followed by Hobq, and then Yumin.

[0097] The station with the weakest spatial heterogeneity of snow density obtained by original point observation results is Tacheng, followed by Qinghe, and then Beita Mountain. The station with the strongest spatial heterogeneity is Jimunai, followed by Hobq, and then Yumin.

[0098] Although the spatial heterogeneity indices of snow cover parameters calculated using the profile average and the original point observations are not entirely consistent, the trends are the same. The higher heterogeneity index obtained from the profile average indicates that the snow cover parameters exhibit significant spatial heterogeneity at larger scales (e.g., km scale), while the higher heterogeneity index obtained from the original observations indicates that the snow cover parameters exhibit significant spatial heterogeneity at smaller spatial scales (e.g., m scale).

[0099] The spatial heterogeneity of snow depth, snowmelt equivalent, and snow density also varies between different stations.

[0100] The results are similar. The station with the highest snow depth heterogeneity index does not necessarily have the highest snowmelt equivalent heterogeneity index, and similarly, the station with the highest snowmelt equivalent spatial heterogeneity index does not necessarily have the highest snow density spatial heterogeneity index. These results indicate that for snow parameter inversion, it is necessary to conduct a targeted evaluation of the spatial representativeness of snow stations for the specific snow parameters to be inverted. Algorithms that directly invert snow depth and calculate snowmelt equivalent using snow density lookup tables need to incorporate the spatial representativeness of snow depth observed at meteorological stations into the algorithm, using the quality indicators of the station's snow depth observation results to evaluate and weight the station data entering the algorithm. Similarly, for snowmelt equivalent algorithms that assimilate snowmelt equivalent data using station snow density data, it is necessary to incorporate the spatial representativeness of snow density observed at meteorological stations into the algorithm, using the quality indicators of the station's snow density observation results to weight and evaluate the station observation data entering the algorithm.

[0101] The weights of the spatial heterogeneity of the aforementioned snow cover parameters can be obtained through ground-based observation experiments or through collaboration with meteorological stations and routine, intensive spatial observation work.

[0102] In addition to the analysis of spatial heterogeneity in snow cover distribution around meteorological stations, we also conducted this ground observation...

[0103] The data obtained from the test can also be used to assess the reliability of the meteorological station's location in snow cover parameter observation. In other words, given the inherent heterogeneity of snow cover spatial distribution around the meteorological station, to what extent can the observations from the station's location more accurately describe the large-scale average distribution of snow cover parameters?

[0104] We defined three sets of indicators to evaluate the snow depth observation data of the meteorological station.

[0105] Absolute deviation of snow depth at weather stations: SD AbDe =SD Station -SD Avg

[0106] Relative deviation of snow depth at weather stations: SDReDe = (SD Station - SD Avg ) / SD Avg

[0107] Snow depth weather station location index:

[0108] According to the above index, the snow depth observation results of the test are calculated, and the calculation results are shown in Figure 4 , wherein the subscripts l and p correspond to the profile mean and the original point, respectively.

[0109] Among the 9 weather observation stations in northern Xinjiang, the station with the smallest relative deviation of snow depth station observation obtained by the profile mean is Altai, followed by Qinghe, and then Beitashan. The relative deviation is the largest in Jimunai, followed by Yumin, and then Tuoli. The station with the smallest relative deviation of snow depth station observation obtained by the original point is Qinghe, followed by Beitashan and Altai. The relative deviation is the largest in Jimunai, followed by Yumin, and then Tuoli.

[0110] Since the snow observation site in the weather station has a certain range, it can be considered that the site selection of the snow observation site can be compared with the profile mean data. Here we use the profile mean to compare when calculating the snow depth station location index.

[0111] Among the 9 weather observation stations in northern Xinjiang, the station with the optimal snow depth station location index obtained by the profile mean is Altai, followed by Beitashan, and then Qinghe. The station with the worst station location index is Yumin, followed by Habah, and then Taicheng.

[0112] As can be seen from the results, although the natural conditions of some stations are good, and the natural distribution of snow in the region is relatively uniform, due to the problem of station location setting, the spatial representativeness is weakened to a certain extent. Although the natural conditions of some stations are poor, and the natural distribution of snow in the region is not uniform, due to the good station location setting, the spatial representativeness is improved to a certain extent.

[0113] In the actual use of station data, it is necessary to distinguish the difference between the natural snow state distribution around the station and the spatial representativeness difference of the station. The former focuses on the mixed pixel problem in the development of snow algorithm, and the latter focuses on the precision problem in the verification of snow depth.

[0114] In this application, we mainly focus on the influence of wind speed and direction on snow redistribution, and the influence on the spatial representativeness of snow parameter observation results of weather stations.

[0115] First, we analyze the impact of wind speed and direction on snow from the ground snow profile observations. In this application, the deepest profile observed is from Jimunai, with two profiles of 75 cm and 120 cm. The average snow depth in Jimunai is only 18.91 cm, which is the smallest among the nine sites involved in our application.

[0116] The snow density of each layer of the profile is much higher than the average snow density around the Jimunai site, and also higher than the average snow density of all meteorological sites involved in this application. Most of the layered snow does not show the characteristic of increasing snow density with increasing depth. This indicates that the abnormal snow density of the snow layer is not caused by the gravity of the snow itself. Similarly, the snow water content does not show the characteristic of gradually increasing with increasing snow depth. The water content is basically irregular except for the larger bottom layer.

[0117] After the profile is dug out, we find that it is located on the windward slope of a small soil bag. Long-term wind blowing snow accumulation has greatly reduced the surface slope, making it difficult to see the terrain features of the underlying surface.

[0118] The profile has a very obvious stratification feature. Unlike other normal stratification features of snow, the stratification of this profile is marked by a clear dust layer, while normal snow stratification should be marked by changes in snow particle size and snow density caused by snow freeze-thaw and natural compaction. In addition to the dust layer near the surface layer, there is also a clear dust stratification in the lower layer of snow. Combined with the local meteorological and geomorphological conditions, it can be basically determined that it is caused by the dust raised after the surface is exposed by wind blowing snow.

[0119] To further analyze the impact of wind speed on the distribution of snow, we extracted the wind speed data of Beita Mountain, Qinghe, Fuyun, Alatai, Habar River, Jimunai, Tuoli, Yumin and Ta Cheng from 1951 to 2015 in December, January and February for analysis (see Figure 5 ):

[0120] The maximum wind speed of the three months of the five sites of Habar River, Jimunai, Tuoli, Yumin and Ta Cheng is significantly higher than that of Beita Mountain, Qinghe, Fuyun, Alatai. Among them, the maximum wind speed of Habar River is the highest, with a maximum wind speed of nearly 13 m / s in January.

[0121] In addition to the maximum wind speed, we also analyzed the average wind speed.

[0122] The above results are basically consistent with the observation data obtained by the method of the application, and only the relationship between the snow observation results of Beita Mountain and the average wind speed is difficult to explain. Further analysis of the correlation between wind speed and snow redistribution is needed for Beita Mountain. In addition to wind speed, wind direction also has a significant impact on snow redistribution. When the wind direction is relatively fixed, the heterogeneity of snow redistribution will be more obvious.

[0123] The above is only the preferred embodiment of the application, which is used to help understand the method of the application and its core idea. The protection scope of the application is not limited to the above-mentioned embodiments. Any technical solution that belongs to the idea of the application is within the protection scope of the application. It should be noted that for ordinary technical personnel in the technical field, some improvements and decorations without departing from the principles of the application should also be considered as the protection scope of the application.

[0124] The application solves the problem of large error in the algorithm accuracy of the snow depth and snow water equivalent algorithm of the satellite microwave imager in the prior art as a whole. Through observation and analysis of important meteorological station snow parameters in the algorithm, the application promotes the work of snow parameter inversion algorithm, snow product verification, snow depth, snow water equivalent business service, and the like.

Claims

1. A method for observing and analyzing the snow parameters around a weather station, characterized in that, The observation results include snow profile observation, UAV observation, summary analysis of observation results, and analysis of spatial distribution heterogeneity of snow, specifically: A. Snow profile observation: obtain the spatial distribution characteristics of snow depth, snow water equivalent, and snow particle size parameters, including snow depth profile observation and snow depth, snow water equivalent, and snow particle size profile observation. The snow depth profile observation is used to obtain the spatial distribution of snow depth to the greatest extent. The snow depth, snow water equivalent, and snow particle size profile observation is based on the results of snow depth spatial distribution, combined with the observation of snow water equivalent and snow particle size, and the results are extended to a 25km scale range. B. Snow profile observation: includes snow density, snow water content, and layered conditions. Except for obvious natural layering, which requires additional observation layering, the rest of the observation starts from the top layer of snow, with snow characteristics observed every 5cm until the ground. C. UAV observation: uses a UAV to take aerial photographs of snow-covered ground. The results of each area flight are spliced to obtain the overall observation results, obtain DEM digital elevation matrix, DOM digital orthographic image, and DSM digital surface model results, and realize the scale conversion from point to plane. D. Summary analysis of observation results: includes analysis of spatial distribution heterogeneity of snow around the weather station and spatial representativeness analysis of the weather station: The spatial distribution heterogeneity of snow around the weather station is quantitatively estimated by the ratio of the standard deviation of the spatial distribution of snow depth, snow water equivalent, and snow density to the average value of the observation: Snow depth spatial heterogeneity index: SD Het = SD Stdev / SD Avg Snow water equivalent spatial heterogeneity index: SWE Het = SWE Stdev / SWE Avg Snow density spatial heterogeneity index: Den Het = Den Stdev / Den Avg Het is the spatial heterogeneity index of snow; According to the spatial heterogeneity index results: For snow parameter inversion, the spatial representativeness of the snow station needs to be evaluated for the specific snow parameter to be inverted. For direct inversion of snow depth, the algorithm for calculating snow water equivalent using snow density lookup table is used to incorporate the spatial representativeness of the weather station's snow depth observation into the algorithm, and the quality identification weight of the station's snow depth observation results is used to evaluate and weight the station data entering the algorithm. For the snow water equivalent algorithm using station snow density data for assimilation, the spatial representativeness of the weather station's snow density observation is incorporated into the algorithm, and the quality identification weight of the station's snow density observation results is used to evaluate and weight the station observation data entering the algorithm. The weight is obtained through ground observation tests or through collaboration with the weather station through daily spatial encryption observation work; Spatial representativeness analysis of the weather station: evaluate the reliability of the weather station's location in terms of snow parameter observation. The observation results of the weather station's location accurately describe the large-scale distribution average of snow parameters, including three groups of indicators to evaluate the snow depth observation data of the weather station: Absolute bias of the snow depth weather station: SD AbDe = SD Station - SD Avg Relative bias of the snow depth weather station: SD ReDe = (SD Station - SD Avg ) / SD Avg Snow depth weather station location index: According to the calculation of the three groups of indicators on the snow depth observation results, according to the calculation results, in the actual use of station data, the difference between the natural snow state distribution around the station and the difference in the spatial representativeness of the station is distinguished. The distribution difference of natural snow state around the site mainly considers the mixed pixel problem in the development of snow algorithm; The spatial representativeness difference of the site mainly considers the precision problem in the verification of snow depth; E. Analysis of spatial distribution heterogeneity of snow: Including the influence of wind speed and direction on snow redistribution and the influence on the spatial representativeness of snow parameter observation results of meteorological stations; From the observation of the influence of wind speed and direction on snow, it is analyzed that the snow density of each layer of the profile is higher than the average value of the snow density around the site, and most of the layered snow does not show the characteristics of increasing snow density with increasing depth, which shows that the abnormal snow density of the profile snow layer is not caused by the gravity of the snow itself, and the snow water content does not show the characteristics of gradually increasing with the increase of snow depth, and the influence of wind blowing snow on snow redistribution is obtained by comparing the profile layering characteristics with other snow layering characteristics; Through the relationship between the maximum and average values of the wind speed of each station and the snow parameters and the relationship between the stationarity of the wind direction and the snow parameters, it is obtained that the wind speed and direction influence the snow redistribution, and further influence the spatial representativeness of the snow parameter observation results of meteorological stations.

Images