Microwave radiometry experiment for snow in Altay 1 China: time series of in situ data for electromagnetic and 2 physical features of snow pack 3

. Snow depth is a key parameter in climatic and hydrological systems. Passive microwave 19 remote sensing, snow process model and data assimilation are the main methods to estimate snow depth 20 in large scale. The estimation accuracies strongly depend on input of snow parameters or characteristics. 21 Because the evolving processes of snow parameters vary spatiotemporally, and are difficult to accurately 22 simulate or observe, large uncertainties and inconsistence exist among existing snow depth products. 23 Therefore, a comprehensive experiment is needed to understand the evolution processes of snow 24 characteristics and their influence on microwave radiation of snowpack, to evaluate and improve the 25 snow depth and SWE retrieval and simulation methods. An Integrated Microwave Radiometry Campaign 26 for snow (IMCS) was conducted at the Altay National Reference Meteorological station (ANRMS) in 27 Xinjiang, China, during snow season of 2015/2016. The campaign hosted a dual polarized microwave 28 radiometer operating at L, K and Ka bands to provide minutely passive microwave observations of snow 29 cover at a fixed site, daily manual snow pit measurements, ten-minute automatic 4-component radiation and layered snow temperatures, covering a full snow season of 2015/2016. The measurements of meteorological and underlying soil parameters were requested from the ANRMS. This study provides a summary of the obtained data, detailing measurement protocols for microwave radiometry, in situ snow pit and station observation data. A brief analysis of the microwave signatures against snow parameters is presented. A consolidated dataset of observations, comprising the ground passive microwave brightness temperatures, in situ snow characteristics, 4-component radiation and weather parameters, was achieved at the National Tibetan Plateau Data Center, China. The dataset is unique in providing continuous daily snow pits data and coincident microwave brightness temperatures, radiation and meteorological data, at a fixed site over a full season. The dataset is expected to serve the evaluation and development of radiative transfer models and snow process models. The consolidated data are available at


21
Because the evolving processes of snow parameters vary spatiotemporally, and are difficult to accurately 22 simulate or observe, large uncertainties and inconsistence exist among existing snow depth products.  neighboring bare rectangle fields in the ANRMS with areas of 2500m 2 (black rectangle filed in Figure   126 1), 2500m 2 (pink rectangle field in Figure 1), 200m 2 (red rectangle field in Figure 1) and 400 m 2 (blue 127 rectangle field in Figure 1), respectively.
only collected layered snow temperatures and 4-component radiation.

144
Because the four observation fields are located within the domain of the station and the distance 145 between them are less than 100m, the snow characteristics and soil and weather conditions are thought 146 to be the same. Overall, the experiment performed a systematic observation covering electromagnetic 147 and physical features of snow pack, providing data for studies on snow remote sensing and models.

153
The blue rectangle is the field for meteorological and soil data collection operated by the ANRMS.    Table 3). Although the snow temperatures were manually measured at snow 200 pits, the automatically collected snow temperatures in the red field were utilized in this study, because 201 the temperature measured at snow pits could not reflect the natural temperature profile when the snow 202 pits exposed to air. The first step of snow pit measurement is making a snow pit. In the black field, a new snow pit was 208 dug each day. A spade was used to excavate snow pit. The length of the snow pit profile was 209 approximately 2m to make sure all parameters were measured from unbroken snowpack. The width of 210 the snow pit was approximately 1m. The snow pit section was made as flat as possible using a flat shovel 211 or ruler. When the snow profile is exposed to air for a long time, the snow characteristics will be 212 influenced by environment and will be different from the natural snow characteristics. In order to make 213 sure every observation conducted on natural snow pit, the snow pit was backfilled with the shoveled 214 snow after finishing all observations, and the new snow pit in the following day was made at least 1-m visually determined, and the thickness of each layer was measured using a ruler.

217
The third step was measuring grain size and shape type in each layer. The grain size and type within 218 each natural layer were estimated visually from a microscope with an "Anyty V500IR/UV" camera 219 ( Figure 3a). A software "VIEWTER Plus" matched the microscope was used to measure grain size. The  Figure A1 presents an example of the original photos of grains in each layer, and Table A1 shows

238
Snow density was measured using three instruments: snow tube, snow shovel and Snow Fork

263
The NR01 net radiometer was set up to measure the energy balance between incoming short-wave 264 and long-wave far infrared radiation versus surface-reflected short-wave and outgoing long-wave 265 radiation. The range of short wave is 285~3000nm, and the range of long wave is 4.5~40um. The 4-266 component radiation was automatically recorded every ten minutes. In addition, the sensor is equipped 267 with a Pt100 temperature sensor for parallel recording of the sensor temperature.

281
The air temperature, pressure and humidity were collected using temperature and wetness sensor in 282 thermometer screen, the wind speed and direction were measured using wind sensor set up at 10 m on a 283 tower. Soil moisture and temperature were automatically measured using moisture sensor and 284 temperature sensor. Figure 6 depicts the instruments for these observations.

323
Hourly weather data include date, hour, air temperature, pressure, humidity, wind speed, soil temperature 324 at 5 cm, 10 cm, 15 cm and 20 cm, and soil moisture at 10 cm and 20 cm.

336
During this snow season, there were seven snowfall events and each formed a distinct snow layer except 337 for the third event whose layering became indistinguishable from the second layer (Figure 7 gray). The

338
fourth event was the biggest, after which time snow depth started to decrease and snow density increased.

339
Snow cover began melting on March 14 and snow depth declined to zero within 10 days.  The soil temperature at 5 and 10 cm remained stable and below 0 o C during the snow season but 384 presented large fluctuation before (after) snow on (off) (Figure 11). The temperature difference between 385 5 cm and 10 cm was much larger before snow cover onset than during snow cover period. The soil 386 moistures at 10 cm were above 10% before snow cover onset and after snow off, and there were two soil 387 moisture peaks, one from December 12-14 and another from January 1-20, within the snow cover period.

392
The microwave brightness temperatures varied with snow and soil characteristics, and weather 393 conditions. Figure 12 shows the daily brightness temperatures, brightness temperature difference 394 between 18 and 36 GHz, and snow depth at 1:00 am local time. Figure 13 shows the hourly variation in

407
The variation of L band was mainly influenced by soil moisture and soil temperature. We have soil 408 temperatures at 0 cm, 5 cm and 10 cm and soil moisture at 10 cm. However, the L band reflects the soil 409 moisture within 5 cm which was absent in this experiment. Actually, we did not find the variation of 410 brightness temperature at L band had relationship with soil moisture at 10 cm and soil temperature.

455
The snow pit data and microwave brightness temperatures have proven useful for evaluating and 456 updating a microwave emission transfer model of snowpack (Dai et al., 2022). This dataset reflected the 457 general fact that brightness temperature at higher frequencies presented stronger volume scattering of 458 snow grains, and were more sensitive to snow characteristics. This experiment revealed that the dominant 459 control for the variation of brightness temperature was the variation of grain size but not the snow depth. temperatures were expected to reflect soil moisture variation which influence the soil transmissivity the stable period. The data can be used to analyzes the evolution process of snow characteristics 520 combining with weather data, validate and improve the snow process models, such as SNOWPACK 521 (Lehning et al., 2002), SNTHERM (Chen et al., 2020). The improvement of these models can further 522 enhance the prediction accuracy of land surface process and hydrology models, and the simulation 523 accuracy of snow microwave emission models.

524
Microwave radiometer data and snow pit data have been utilized to analyze the volume scattering 525 features of snow pack at different frequencies (Dai et al., 2022). Results showed that grain size is the 526 most important factor to influence snow volume scattering. The data can also be used to further analyze 527 polarization characteristics of snow pack combining with soil and weather data, and be used to validate 528 different microwave emission models of snowpack.

529
The microwave and optical radiations were simultaneously observed. Existing studies reported that

547
Competing interests: The authors declare that they have no conflict of interest.

549
Acknowledgment: The authors would like to thank the Altay meteorological station for providing    References: Validation of the SNTHERM Model Applied for Snow Depth, Grain Size, and Brightness Temperature