本幅2010年青藏高原多年冻土分布图，以遥感冻结/融化指数为驱动，以高质量区域调查图为约束，采用了一个全新的制图方法。制图方法中通过一个土壤参数来表现局部环境因素对多年冻土分布的影响，并通过空间聚类以及以高质量的区域调查图为目标对该参数进行率定。本图显示，2010年青藏高原地区多年冻土面积为108.6万平方公里,约占青藏高原总面积的41.2%，季节性冻土面积约为144.7万平方公里。此分布图与调查图相比的kappa系数为0.74；以钻孔数据为验证目标，总体精度为0.85，kappa系数为0.43。可靠的精度使得此冻土分布图可以作为全球变暖背景下青藏高原地区多年冻土模拟的标定基准和历史参考。数据格式为geotiff，空间分辨率约1km，投影为经纬度。
", "ds_source": "1.地表温度数据
\n美国国家航空航天局提供每日MODIS LST/发射率产品（MOD11A1和MYD11A1版本6）和NDVI产品（MOD13A2），下载于https://www.earthdata.nasa.gov/.
\n2.DEM数据
\n航天飞机雷达地形任务90m数字高程数据库（SRTM/DEM，第4版）下载于https://cgiarcsi.community/data/srtm-90m-digital-elevation-database-v4-1/.
\n3.降水量数据
\n中国1公里月降水量数据集 (Peng et al., 2019) ，获取于https://doi.org/10.5281/zenodo.3114194.
\n4.积雪数据
\n高亚洲500米日部分积雪数据集(Qiu et al., 2017)，获取于https://doi.org/10.11888/GlaciolGeocryol.tpe.0000016.file.
\n5.土壤属性数据
\n用于地表建模的中国土壤属性数据集(Shangguan et al., 2013) ，获取于http://globalchange.bnu.edu.cn/research/soil2.
\n6.气象站数据
\n中国国家地面气象站（3.0版）由中国国家气象信息中心提供，下载地址为https://data.tpdc.ac.cn/en/data/52c77e9c-df4a-4e27-8e97-d363fdfce10a/.
\n7.钻孔数据
\n青藏高原地区长期冻土检测钻孔地温数据(Zhao et al., 2021) ，获取于https://doi.org/10.11888/Geocry.tpdc.271107.
", "ds_process_way": "1.获取地表冻融指数。
\n基于中分辨率成像光谱仪（MODIS）的地表温度（LST）数据计算地表冻结指数与融化指数。我们采用我们发展的方法来解决LST的缺值问题。首先，采用一种贝叶斯方法对LST缺失值估计等效晴空LST（即假设无云的LST值）[1]。然后，使用了基于太阳-云-卫星几何（SCSG）的云下LST重建方法[2]，从等效晴空LST中消除消除云影响得到云下LST。剩余空缺数据不到10%。计算日均LST，对空缺的日均LST进行常用时空插值得到全年日均LST数据。
\n2.热偏移校正。
\n由于冻土0cm地面温度（GST）会受到季节性积雪或植被的影响，GST和遥感LST数据产品之间通常存在热偏移，因此需要对LST进行校正以获得GST，并据此计算地表冻融指数作为冻土赋存模拟的驱动数据。我们利用站点观测建立GST和LST、归一化植被指数（NDVI）、纬度之间的回归关系，实现从日均LST到日均GST的估计，并在此基础上计算得到地表冻结指数和融化指数。
\n3.基于调查图的青藏高原多年冻土分布制图。
\nFROSTNUM/COP多年冻土制图方法[3]利用多年冻土区域调查图为基准，使用空间聚类和参数优化来估计土壤参数，这些参数在改进的地表冻结数（FROSTNUM）模型中可以表征与冻土分布相关的土壤条件的空间异质性，从而考虑局地因子对多年冻土分布的影响，提高模拟精度。采用能够同时处理连续变量和类别变量的K-prototypes方法进行土壤的空间聚类，并使用效率较高的粒子群算法作为参数优化方法对土壤参数进行率定。更重要的是，针对FROSTNUM/COP方法在较大区域应用时易出现的异参同效的问题进行改进，在参数优化中引入更多的约束以保证参数优化的稳定性。最终以之前得到的地表冻融指数为驱动数据，利用2010年附近的区域冻土调查图为优化目标，模拟得到2010年青藏高原地区的冻土类型分布。
\n4.验证
\n与钻孔以及现有多年冻土分布图进行对比分析，以验证准确性。
\n[1] Chen Y, Nan Z, Zhao S, et al. A Bayesian Approach for Interpolating Clear-Sky MODIS Land Surface Temperatures on Areas With Extensive Missing Data. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 14: 515-528.
\n[2] Chen Y, Nan Z, Cao Z, et al. A stepwise framework for interpolating land surface temperature under cloudy conditions based on the solar-cloud-satellite geometry. ISPRS Journal of Photogrammetry and Remote Sensing, 2023, 197: 292-308.
\n[3] Hu J, Zhao S, Nan Z, et al. An effective approach for mapping permafrost in a large area using subregion maps and satellite data. Permafrost and Periglacial Processes, 2020, 31(4): 548-560.
", "ds_quality": "数据精度：
\n（1）此分冻土布图与典型区调查图相比的kappa系数为0.74；
\n（2）以钻孔数据为验证目标，总体精度为0.85，kappa系数为0.43
", "ds_acq_start_time": "2010-01-01 00:00:00", "ds_acq_end_time": "2010-12-31 00:00:00", "ds_acq_place": "青藏高原", "ds_acq_lon_east": 104.4, "ds_acq_lat_south": 26.0, "ds_acq_lon_west": 73.4, "ds_acq_lat_north": 39.766666666666666, "ds_acq_alt_low": 8500.0, "ds_acq_alt_high": null, "ds_share_type": "open-access", "ds_total_size": 676207, "ds_files_count": 2, "ds_format": "geotiff", "ds_space_res": "1000", "ds_time_res": "无", "ds_coordinate": "WGS84", "ds_projection": "GCS_WGS_1984", "ds_thumbnail": "3372d42f-7df8-4ee2-868d-4ee92be9c21f.png", "ds_thumb_from": 0, "ds_ref_way": "", "paper_ref_way": "", "ds_ref_instruction": "", "ds_from_station": null, "organization_id": "52b7b79b-860c-49a5-9083-9a70cf8bed5a", "ds_serv_man": "李红星", "ds_serv_phone": "0931-4967592", "ds_serv_mail": "ncdc@lzb.ac.cn", "doi_value": "", "subject_codes": [ "170.4510" ], "quality_level": 3, "publish_time": "2023-09-08 14:59:24", "last_updated": "2023-09-08 14:59:13", "protected": false, "protected_to": null, "lang": "zh", "cstr": "11738.11.NCDC.PERMALAB.DB3965.2023", "i18n": { "en": { "title": "A new permafrost distribution map over the Qinghai-Tibet Plateau for 2010", "ds_format": "geotiff", "ds_source": "1.Landsurface temperature data
\nDaily MODIS LST/emissivity products (MOD11A1 and MYD11A1 version 6) and the NDVI product (MOD13A2) are provided by NASA and available at https://www.earthdata.nasa.gov/.
\n2.DEM data
\nThe Shuttle Radar Topography Mission 90m digital elevation database (SRTM/DEM, version 4) is available at https://cgiarcsi.community/data/srtm-90m-digital-elevation-database-v4-1/.
\n3.Precipitation data
\nThe 1-km monthly precipitation dataset for China (Peng et al., 2019) is available at https://doi.org/10.5281/zenodo.3114194.
\n4.Snow cover data
\nThe 500 m Daily Fractional Snow Cover Dataset Over High Asia (Qiu et al., 2017) is available at https://doi.org/10.11888/GlaciolGeocryol.tpe.0000016.file.
\n5.Soil properties data
\nThe China Data Set of Soil Properties for Land Surface Modeling (Shangguan et al., 2013) is available at http://globalchange.bnu.edu.cn/research/soil2.
\n6.Weather station data
\nThe China national surface weather stations (version 3.0) is provided by China National Meteorological Information Center and available at https://data.tpdc.ac.cn/en/data/52c77e9c-df4a-4e27-8e97-d363fdfce10a/.
\n7.Borehole data
\nThe borehole ground temperature data provided by (Zhao et al., 2021) is available at https://doi.org/10.11888/Geocry.tpdc.271107.
", "ds_quality": "Accuracy of Data：
\n(1) The validations against survey-based subregion permafrost maps showed the Kappa coefficient of 0.74.
\n(2) The validations against borehole records showed the overall Accuracy of 0.85 and Kappa coefficient of 0.43.
", "ds_ref_way": "", "ds_abstract": "This new permafrost distribution map over the Qinghai-Tibet Plateau (QTP) for 2010 was created through a novel permafrost mapping approach with satellite data as inputs and survey-based subregion permafrost maps as constraints. The ground surface thawing and freezing indices, as driving data for the mapping approach, were derived from remotely sensed land surface temperature data. The mapping approach takes the effects of local factors into account by incorporating a permafrost-related soil parameter in the model whose values are optimally estimated through spatial clustering based on environmental variables and parameter optimization technique with the survey-based subregion permafrost maps as an optimization objective. This new map indicates a total permafrost area of about 1.086×106 km2 (41.2% of the QTP) and seasonally frozen ground of about 1.447×106 km2 (54.9% of the QTP) in 2010 with glaciers and lakes excluded. The validations against survey-based subregion permafrost maps (κ = 0.74) and borehole records (Overall Accuracy = 0.85 and κ = 0.43) showed a higher accuracry compared with two other recent maps. This map can serve as a benchmark map at sufficient quality for land surface simulations and as a historical reference for projecting future permafrost changes on the QTP. Data format: geotiff; spatial resolution: approx. 1km; spatial reference: lat/lon (WGS84).
", "ds_time_res": "无", "ds_acq_place": "Qinghai-Tibet Plateau", "ds_space_res": "1000", "ds_projection": "GCS_WGS_1984", "ds_process_way": "Computation of Ground Surface Freeze and Thaw Indices:
\nThe computation of ground surface freeze and thaw indices relied on Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature (LST) data. To address gaps in LST data, we developed a specialized methodology. Initially, we employed a Bayesian approach to estimate clear-sky equivalent LST values (i.e., assuming cloud-free LST) [1]. Subsequently, cloudy LST values were derived by removing cloud effects from the clear-sky equivalent LST, utilizing a solar-cloud-satellite geometry (SCSG)-based method [2]. The remaining missing data comprised less than 10%. Daily mean LST was calculated, and conventional temporal and spatial interpolation techniques were applied to fill in the missing daily mean LST values, yielding annual daily mean LST data.
\nThermal Offset Correction:
\nDue to the influence of seasonal snow cover or vegetation on ground surface temperature (GST) in permafrost regions, a typical thermal offset exists between GST and remotely sensed LST data products. Consequently, LST required correction to obtain GST, which was then employed to compute the ground surface freeze-thaw index for permafrost simulations. A regression relationship was established, utilizing site observations, among GST, LST, normalized vegetation index (NDVI), and latitude. This facilitated the estimation of daily mean GST from daily mean LST, subsequently leading to the computation of surface freeze and thaw indices based on the daily mean GST data.
\nMapping Permafrost Distribution on the Qinghai-Tibet Plateau Using Subregion Survey Maps:
\nThe FROSTNUM/COP method for permafrost mapping [3] employs subregion survey maps as constraints. Spatial clustering and parameter optimization were employed to estimate soil parameters capable of characterizing spatial heterogeneity in soil conditions relevant to permafrost distribution. This approach accounted for the influence of local factors on permafrost distribution, thereby enhancing simulation accuracy. Soil spatial clustering was achieved using the k-prototypes method, which can handle both numerical and categorical variables. Parameter optimization was carried out using the particle swarm algorithm. To address the challenge of parameter equifinality, effective constraints were introduced to ensure efficiency. Ultimately, utilizing previously acquired surface freeze-thaw indices as driving data and 2010 regional permafrost survey maps as optimization targets, we generated a map depicting the distribution of frozen ground types on the Qinghai-Tibet Plateau for the year 2010.
\nVerification:
\nA comparative analysis was conducted, contrasting the results with borehole data and existing permafrost distribution maps to validate accuracy.
\nReferences
\n[1] Chen Y, Nan Z, Zhao S, et al. \"A Bayesian Approach for Interpolating Clear-Sky MODIS Land Surface Temperatures on Areas With Extensive Missing Data.\" IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 14: 515-528.
\n[2] Chen Y, Nan Z, Cao Z, et al. \"A Stepwise Framework for Interpolating Land Surface Temperature under Cloudy Conditions Based on the Solar-Cloud-Satellite Geometry.\" ISPRS Journal of Photogrammetry and Remote Sensing, 2023, 197: 292-308.
\n[3] Hu J, Zhao S, Nan Z, et al. \"An Effective Approach for Mapping Permafrost in a Large Area Using Subregion Maps and Satellite Data.\" Permafrost and Periglacial Processes, 2020, 31(4): 548-560.
", "ds_ref_instruction": "" } }, "submit_center_id": "ncdc", "data_level": 0, "license_type": "CC BY 4.0", "doi_reg_from": "reg_local", "cstr_reg_from": "reg_local", "doi_not_reg_reason": null, "cstr_not_reg_reason": null, "is_paper_in_submitting": false, "ds_topic_tags": [ "青藏高原", "多年冻土分布图", "基准图" ], "ds_subject_tags": [ "自然地理学" ], "ds_class_tags": [], "ds_locus_tags": [ "高亚洲", "青藏高原" ], "ds_time_tags": [ 2010 ], "ds_contributors": [ { "true_name": "南卓铜", "email": "nanzt@shnu.edu.cn", "work_for": "上海师范大学", "country": "中国" } ], "ds_meta_authors": [ { "true_name": "曹泽涛", "email": "cao.zt@outlook.com", "work_for": "南京师范大学地理科学学院", "country": "中国" } ], "ds_managers": [ { "true_name": "南卓铜", "email": "nanzt@shnu.edu.cn", "work_for": "上海师范大学", "country": "中国" } ], "category": "冻土" }