Search

The spatial decomposition of demographic data at a fine resolution is a classic and crucial problem in the field of geographical information science. The main objective of this study was to compare twelve well-known machine learning regression algorithms for the spatial decomposition of demographic data with multisource geospatial data. Grid search and cross-validation methods were used to ensure that the optimal model parameters were obtained. The results showed that all the global regression algorithms used in the study exhibited acceptable results, besides the ordinary least squares (OLS) algorithm. In addition, the regularization method and the subsetting method were both useful for alleviating overfitting in the OLS model, and the former was better than the latter. The more competitive performance of the nonlinear regression algorithms than the linear regression algorithms implies that the relationship between population density and influence factors is likely to be non-linear. Among the global regression algorithms used in the study, the best results were achieved by the k-nearest neighbors (KNN) regression algorithm. In addition, it was found that multi-sources geospatial data can improve the accuracy of spatial decomposition results significantly, and thus the proposed method in our study can be applied to the study of spatial decomposition in other areas.

China is particularly vulnerable to earthquakes and has suffered a high number of earthquake casualties. Therefore, it is imperative to develop effective methods for assessing population vulnerability in order to provide practical and useful information for pre-disaster preparedness and emergency management. However, there are many problems with existing approaches to assessing population vulnerability in China. This study developed a method combining dasymetric population mapping with population vulnerability indexing to estimate populations&rsquo, vulnerability to earthquakes at block level in daytime and nighttime. The method aims to provide high spatial-temporal resolution information on vulnerable populations and population vulnerability. In a case study of Haidian District in Beijing, we combined our population spatialization model with a population vulnerability assessment model and sensitivity analysis to efficiently identify areas with high population vulnerability, the location of vulnerable populations, and key sensitive factors influencing population vulnerability during daytime and nighttime. This information can be used to identify where and which aspects of pre-disaster mitigation, preparedness, and emergency response activities should be prioritized. It is expected that these methods will be used to analyze vulnerable populations and population vulnerability in other areas of China.

Where urban dwellers live at a fine scale is essential for the planning of services and response to city emergencies. Currently, most existing population mapping approaches considered census data as observational data for specifying models. However, census data usually have low spatial resolution and low frequency. Here, we presented a framework for mapping populations in residential neighborhoods with 30 m spatial resolution with little dependency upon census data. The framework integrated remote sensing and crowdsourcing data. The observational populations and number of households at residential neighborhood scale were obtained from real-time crowdsourcing data instead of census data. We tested our framework in Beijing. We found that (1) the number of households from a real estate trade platform could be a good proxy for accurate observational population. (2) The accuracy of the mapping population in residential neighborhoods was reasonable. The mean absolute percentage error was 47.26% and the R2 was 0.78. (3) Our framework shows great potential in mapping the population in real time. Our findings expand the knowledge in estimating urban population. In addition, the proposed framework and approach provide an effective means to quantify population distribution data for cities, which is particularly important for many of the cities worldwide lacking census data at the residential neighborhood scale.

Sea-ice biophysical properties are characterized by high spatio-temporal variability ranging from the meso- to the millimeter scale. Ice coring is a common yet coarse point sampling technique that struggles to capture such variability in a non-invasive manner. This hinders quantification and understanding of ice algae biomass patchiness and its complex interaction with some of its sea ice physical drivers. In response to these limitations, a novel under-ice sled system was designed to capture proxies of biomass together with 3D models of bottom topography of land-fast sea-ice. This system couples a pushbroom hyperspectral imaging (HI) sensor with a standard digital RGB camera and was trialed at Cape Evans, Antarctica. HI aims to quantify per-pixel chlorophyll-a content and other ice algae biological properties at the ice-water interface based on light transmitted through the ice. RGB imagery processed with digital photogrammetry aims to capture under-ice structure and topography. Results from a 20 m transect capturing a 0.61 m wide swath at sub-mm spatial resolution are presented. We outline the technical and logistical approach taken and provide recommendations for future deployments and developments of similar systems. A preliminary transect subsample was processed using both established and novel under-ice bio-optical indices (e.g., normalized difference indexes and the area normalized by the maximal band depth) and explorative analyses (e.g., principal component analyses) to establish proxies of algal biomass. This first deployment of HI and digital photogrammetry under-ice provides a proof-of-concept of a novel methodology capable of delivering non-invasive and highly resolved estimates of ice algal biomass in-situ, together with some of its environmental drivers. Nonetheless, various challenges and limitations remain before our method can be adopted across a range of sea-ice conditions. Our work concludes with suggested solutions to these challenges and proposes further method and system developments for future research.

Over recent decades, fine-scale land use and land cover classification in open-pit mine areas (LCCMA) has become very important for understanding the influence of mining activities on the regional geo-environment, and for environmental impact assessment procedure. This research reviews advances in fine-scale LCCMA from the following aspects. Firstly, it analyzes and proposes classification thematic resolution for LCCMA. Secondly, remote sensing data sources, features, feature selection methods, and classification algorithms for LCCMA are summarized. Thirdly, three major factors that affect LCCMA are discussed: significant three-dimensional terrain features, strong LCCMA feature variability, and homogeneity of spectral-spatial features. Correspondingly, three key scientific issues that limit the accuracy of LCCMA are presented. Finally, several future research directions are discussed: (1) unitization of new sensors, particularly those with stereo survey ability; (2) procurement of sensitive features by new sensors and combinations of sensitive features using novel feature selection methods; (3) development of robust and self-adjusted classification algorithms, such as ensemble learning and deep learning for LCCMA; and (4) application of fine-scale mining information for regularity and management of mines.