Land cover dataset of the China Central-Asia West-Asia Economic Corridor from 1993 to 2018.

Land Cover (LC) maps offer vital knowledge for various studies, ranging from sustainable development to climate change. The China Central-Asia West-Asia Economic Corridor region, as a core component of the Belt and Road initiative program, has been experiencing some of the most severe LC change tragedies, such as the Aral Sea crisis and Lake Urmia shrinkage, in recent decades. Therefore, there is a high demand for producing a fine-resolution, spatially-explicit, and long-term LC dataset for this region. However, except China, such dataset for the rest of the region (Kyrgyzstan, Turkmenistan, Kazakhstan, Uzbekistan, Tajikistan, Turkey, and Iran) is currently lacking. Here, we constructed a historical set of six 30-m resolution LC maps between 1993 and 2018 at 5-year time intervals for the seven countries where nearly 200,000 Landsat scenes were classified into nine LC types within Google Earth Engine cloud computing platform. The generated LC maps displayed high accuracies. This publicly available dataset has the potential to be broadly applied in environmental policy and management.

Remote Sensing Image Classification with a Graph-Based Pre-Trained Neighborhood Spatial Relationship.

Previous knowledge of the possible spatial relationships between land cover types is one factor that makes remote sensing image classification "smarter". In recent years, knowledge graphs, which are based on a graph data structure, have been studied in the community of remote sensing for their ability to build extensible relationships between geographic entities. This paper implements a classification scheme considering the neighborhood relationship of land cover by extracting information from a graph. First, a graph representing the spatial relationships of land cover types was built based on an existing land cover map. Empirical probability distributions of the spatial relationships were then extracted using this graph. Second, an image was classified based on an object-based fuzzy classifier. Finally, the membership of objects and the attributes of their neighborhood objects were joined to decide the final classes. Two experiments were implemented. Overall accuracy of the two experiments increased by 5.2% and 0.6%, showing that this method has the ability to correct misclassified patches using the spatial relationship between geo-entities. However, two issues must be considered when applying spatial relationships to image classification. The first is the "siphonic effect" produced by neighborhood patches. Second, the use of global spatial relationships derived from a pre-trained graph loses local spatial relationship in-formation to some degree.

Assessment of five satellite-derived LAI datasets for GPP estimations through ecosystem models.

Ecosystem models have been widely used for obtaining gross primary productivity (GPP) estimations at multiple scales. Leaf area index (LAI) is a critical variable in these models for describing the vegetation canopy structure and predicting vegetation-atmosphere interactions. However, the uncertainties in LAI datasets and the effects of their representation on simulated GPP remain unclear, especially over complex terrain. Here, five most popular datasets, namely the Long-term Global Mapping (GLOBMAP) LAI, Global LAnd Surface Satellite (GLASS) LAI, Geoland2 version 1 (GEOV1) LAI, Global Inventory Monitoring and Modeling System (GIMMS) LAI, and Moderate Resolution Imaging Spectroradiometer (MODIS) LAI, were selected to examine the influences of LAI representation on GPP estimations at 95 eddy covariance (EC) sites. The GPP estimations from the Boreal Ecosystem Productivity Simulator (BEPS) model and the Eddy Covariance Light Use Efficiency (EC-LUE) model were evaluated against EC GPP to assess the performances of LAI datasets. Results showed that MODIS LAI had stronger linear correlations with GLASS and GEOV1 than GIMMS and GLOMAP at the study sites. The GPP estimations from GLASS LAI had a better agreement with EC GPP than those from other four LAI datasets at forest sites, while the GPP estimations from GEOVI LAI matched best with EC GPP at grass sites. Additionally, the GPP estimations from GLASS and GEOVI LAI presented better performances than the other three LAI datasets at crop sites. Besides, the results also showed that complex terrain had larger discrepancies of LAI and GPP estimations, and flat terrain presented better performances of LAI datasets in GPP estimations. Moreover, the simulated GPP from BEPS was more sensitive to LAI than those from EC - LUE, suggesting that LAI datasets can also lead to different uncertainties in GPP estimations from different model structures. Our study highlights that the satellite-derived LAI datasets can cause uncertainties in GPP estimations through ecosystem models.

[An Improved DDV Method to Retrieve AOT for HJ CCD Image in Typical Mountainous Areas].

Domestic HJ CCD imaging applications in environment and disaster monitoring and prediction has great potential. But, HJ CCD image lack of Mid-Nir band can not directly retrieve Aerosol Optical Thickness (AOT) by the traditional Dark Dense Vegetation (DDV) method, and the mountain AOT changes in space-time dramatically affected by the mountain environment, which reduces the accuracy of atmospheric correction. Based on wide distribution of mountainous dark dense forest, the red band histogram threshold method was introduced to identify the mountainous DDV pixels. Subsequently, the AOT of DDV pixels were retrieved by lookup table constructed by 6S radiative transfer model with assumption of constant ratio between surface reflectance in red and blue bands, and then were interpolated to whole image. MODIS aerosol product and the retrieved AOT by the proposed algorithm had very good consistency in spatial distribution, and HJ CCD image was more suitable for the remote sensing monitoring of aerosol in mountain areas, which had higher spatial resolution. Their fitting curve of scatterplot was y = 0.828 6x-0.01 and R2 was 0.984 3 respectively. Which indicate the improved DDV method can effectively retrieve AOT, and its precision can satisfy the atmospheric correction and terrain radiation correction for Hj CCD image in mountainous areas. The improvement of traditional DDV method can effectively solve the insufficient information problem of the HJ CCD image which have only visible light and near infrared band, when solving radiative transfer equation. Meanwhile, the improved method fully considered the influence of mountainous terrain environment. It lays a solid foundation for the HJ CCD image atmospheric correction in the mountainous areas, and offers the possibility for its automated processing. In addition, the red band histogram threshold method was better than NDVI method to identify mountain DDV pixels. And, the lookup table and ratio between surface reflectance between red and blue bands were the important influence factor for AOT retrieval. These will be the important research directions to further improve algorithm and improve the retrieve accuracy.