New Perspectives for Mapping Global Population Distribution Using World Settlement Footprint Products

Spatially explicit accuracy assessment of deep learning-based, fine-resolution built-up land data in the United States

Geospatial datasets derived from remote sensing data by means of machine learning methods are often based on probabilistic outputs of abstract nature, which are difficult to translate into interpretable measures. For example, the Global Human Settlement Layer GHS-BUILT-S2 product reports the probability of the presence of built-up areas in 2018 in a global 10 m × 10 m grid. However, practitioners typically require interpretable measures such as binary surfaces indicating the presence or absence of built-up areas or estimates of sub-pixel built-up surface fractions. Herein, we assess the relationship between the built-up probability in GHS-BUILT-S2 and reference built-up surface fractions derived from a highly reliable reference database for several regions in the United States. Furthermore, we identify a binarization threshold using an agreement maximization method that creates binary built-up land data from these built-up probabilities. These binary surfaces are input to a spatially explicit, scale-sensitive accuracy assessment which includes the use of a novel, visual-analytical tool which we call focal precision-recall signature plots. Our analysis reveals that a threshold of 0.5 applied to GHS-BUILT-S2 maximizes the agreement with binarized built-up land data derived from the reference built-up area fraction. We find high levels of accuracy (i.e., county-level F-1 scores of almost 0.8 on average) in the derived built-up areas, and consistently high accuracy along the rural-urban gradient in our study area. These results reveal considerable accuracy improvements in human settlement models based on Sentinel-2 data and deep learning, as compared to earlier, Landsat-based versions of the Global Human Settlement Layer.

High-resolution gridded population datasets for Latin America and the Caribbean using official statistics

"Leaving no one behind" is the fundamental objective of the 2030 Agenda for Sustainable Development. Latin America and the Caribbean is marked by social inequalities, whilst its total population is projected to increase to almost 760 million by 2050. In this context, contemporary and spatially detailed datasets that accurately capture the distribution of residential population are critical to appropriately inform and support environmental, health, and developmental applications at subnational levels. Existing datasets are under-utilised by governments due to the non-alignment with their own statistics. Therefore, official statistics at the finest level of administrative units available have been implemented to construct an open-access repository of high-resolution gridded population datasets for 40 countries in Latin American and the Caribbean. These datasets are detailed here, alongside the 'top-down' approach and methods to generate and validate them. Population distribution datasets for each country were created at a resolution of 3 arc-seconds (approximately 100 m at the equator), and are all available from the WorldPop Data Repository.

Global 30 meters spatiotemporal 3D urban expansion dataset from 1990 to 2010

Understanding the spatiotemporal dynamics of global 3D urban expansion over time is becoming increasingly crucial for achieving long-term development goals. In this study, we generated a global dataset of annual urban 3D expansion (1990-2010) using World Settlement Footprint 2015 data, GAIA data, and ALOS AW3D30 data with a three-step technical framework: (1) extracting the global constructed land to generate the research area, (2) neighborhood analysis to calculate the original normalized DSM and slope height of each pixel in the study area, and (3) slope correction for areas with a slope greater than 10° to improve the accuracy of estimated building heights. The cross-validation results indicate that our dataset is reliable in the United States(R² = 0.821), Europe(R² = 0.863), China(R² = 0.796), and across the world(R² = 0.811). As we know, this is the first 30-meter 3D urban expansion dataset across the globe, which can give unique information to understand and address the implications of urbanization on food security, biodiversity, climate change, and public well-being and health.

Long-term exposure and health risk assessment from air pollution: impact of regional scale mobility

BACKGROUND

The negative effect of air pollution on human health is widely reported in recent literature. It typically involves urbanized areas where the population is concentrated and where most primary air pollutants are produced. A comprehensive health risk assessment is therefore of strategic importance for health authorities.

METHODS

In this study we propose a methodology to perform an indirect and retrospective health risk assessment of all-cause mortality associated with long-term exposure to particulate matter less than 2.5 microns (PM2.5), nitrogen dioxide (NO2) and ozone (O3) in a typical Monday to Friday working week. A combination of satellite-based settlement data, model-based air pollution data, land use, demographics and regional scale mobility, allowed to examine the effect of population mobility and pollutants daily variations on the health risk. A Health Risk Increase (HRI) metric was derived on the basis of three components: hazard, exposure and vulnerability, utilizing the relative risk values from the World Health Organization. An additional metric, the Health Burden (HB) was formulated, which accounts for the total number of people exposed to a certain risk level.

RESULTS

The effect of regional mobility patterns on the HRI metric was assessed, resulting in an increased HRI associated with all three stressors when considering a dynamic population compared to a static one. The effect of diurnal variation of pollutants was only observed for NO2 and O3. For both, the HRI metric resulted in significantly higher values during night. Concerning the HB parameter, we identified the commuting flows of the population as the main driver in the resulting metric.

CONCLUSIONS

This indirect exposure assessment methodology provides tools to support policy makers and health authorities in planning intervention and mitigation measures. The study was carried out in Lombardy, Italy, one of the most polluted regions in Europe, but the incorporation of satellite data makes our approach valuable for studying global health.

An assessment of people living by coral reefs over space and time

Human populations near ecosystems are used as both a proxy for dependency on ecosystems, and conversely to estimate threats. Consequently, the number of people living near coral reefs is often used in regional coral reef management, evaluation of risk at regional and global scales, and even considerations of funding needs. Human populations and their statistics, are ever-changing and data relating to coral reefs have not been updated regularly. Here, we present an up-to-date analysis of the abundance, and density of people living within 5-100 km of coral reef ecosystems along with population proportion, using freely available data sets and replicable methods. We present trends of changes in human populations living near coral reefs over a 20-year time period (2000-2020), divided by region and country, along with socio-economic denominations such as country income category and Small Island Developing States (SIDS). We find that across 117 coral reef countries there are currently close to a billion people living within 100 km of a coral reef (~13% of the global population) compared with 762 million people in 2000. Population growth by coral reefs is higher than global averages. The Indian Ocean saw a 33% increase in populations within 100 km of a coral reef and 71% at 5 km. There are 60 countries with 100% of their population within 100 km of coral reefs. In SIDS, the proportion of the total population within 100 km of a coral reef is extremely high: 94% in 2020. Population density 5-10 km from coral reefs is 4× the global average. From 5 to 100 km, more people from lower-middle-income countries live by coral reefs than any other income category. Our findings provide the most up-to-date and extensive statistics on the regional and nation-level differences in population trends that play a large role in coral reef health and survival.

Empiric recommendations for population disaggregation under different data scenarios

High-resolution population mapping is of high relevance for developing and implementing tailored actions in several fields: From decision making in crisis management to urban planning. Earth Observation has considerably contributed to the development of methods for disaggregating population figures with higher resolution data into fine-grained population maps. However, which method is most suitable on the basis of the available data, and how the spatial units and accuracy metrics affect the validation process is not fully known. We aim to provide recommendations to researches that attempt to produce high-resolution population maps using remote sensing and geospatial information in heterogeneous urban landscapes. For this purpose, we performed a comprehensive experimental research on population disaggregation methods with thirty-six different scenarios. We combined five different top-down methods (from basic to complex, i.e., binary and categorical dasymetric, statistical, and binary and categorical hybrid approaches) on different subsets of data with diverse resolutions and degrees of availability (poor, average and rich). Then, the resulting population maps were systematically validated with a two-fold approach using six accuracy metrics. We found that when only using remotely sensed data the combination of statistical and dasymetric methods provide better results, while highly-resolved data require simpler methods. Besides, the use of at least three relative accuracy metrics is highly encouraged since the validation depends on level and method. We also analysed the behaviour of relative errors and how they are affected by the heterogeneity of the urban landscape. We hope that our recommendations save additional efforts and time in future population mapping.

Differences between gridded population data impact measures of geographic access to healthcare in sub-Saharan Africa

Background

Access to healthcare is imperative to health equity and well-being. Geographic access to healthcare can be modeled using spatial datasets on local context, together with the distribution of existing health facilities and populations. Several population datasets are currently available, but their impact on accessibility analyses is unknown. In this study, we model the geographic accessibility of public health facilities at 100-meter resolution in sub-Saharan Africa and evaluate six of the most popular gridded population datasets for their impact on coverage statistics at different administrative levels.

Methods

Travel time to nearest health facilities was calculated by overlaying health facility coordinates on top of a friction raster accounting for roads, landcover, and physical barriers. We then intersected six different gridded population datasets with our travel time estimates to determine accessibility coverages within various travel time thresholds (i.e., 30, 60, 90, 120, 150, and 180-min).

Results

Here we show that differences in accessibility coverage can exceed 70% at the sub-national level, based on a one-hour travel time threshold. The differences are most notable in large and sparsely populated administrative units and dramatically shape patterns of healthcare accessibility at national and sub-national levels.

Conclusions

The results of this study show how valuable and critical a comparative analysis between population datasets is for the derivation of coverage statistics that inform local policies and monitor global targets. Large differences exist between the datasets and the results underscore an essential source of uncertainty in accessibility analyses that should be systematically assessed.

Biogas Production Enhancement through Chicken Manure Co-Digestion with Pig Fat

Chicken manure and pig fat are found abundantly around the globe, and there is a challenge to get rid of them. This waste has considerable energy potential to be recovered into fuel, but extracting this energy from some by-products, especially fat, isn’t an easy task. When anaerobic digestion technology stepped to the level of anaerobic co-digestion, the utilisation of hardly degradable waste became feasible. Our research was conducted on anaerobic co-digestion of chicken manure as the primary substrate with pig fat as a fat reach supplement in a semi-continuous mode at different organic load rates. The influence of fat waste on the process of biogas production from chicken manure and the composition of the obtained products was determined using an organic load rate of 3.0–4.5 kg VS·(m3·day)−1. A sturdy and continuously growing biogas production was observed at all organic load rates, implying the synergetic effect on chicken manure and pig fat co-digestion. The highest specific methane yield, 441.3 ± 7.6 L·kg VS−1, was observed at an organic load rate of 4.5 kg VS·(m3·day)−1. The research results showed that co-digestion of chicken manure with pig fat is an appropriate measure for fat utilisation and contributes to the increase in biogas yield, methane concentration, and overall methane yield at investigated organic load rates.

The Ratio of the Land Consumption Rate to the Population Growth Rate: A Framework for the Achievement of the Spatiotemporal Pattern in Poland and Lithuania

Indicator 11.3. 1 of the 2030 sustainable development goals (SDG) 11, i.e., the ratio of the land use to the population growth rate, is currently classified by the United Nations as a Tier II indicator, as there is a globally-accepted methodology for its calculation, but the data are not available, nor are not regularly updated. Recently, the increased availability of remotely sensed data and products allows not only for the calculation of the SDG 11.3. 1, but also for its monitoring at different levels of detail. That is why this study aims to address the interrelationships between population development and land use changes in Poland and Lithuania, two neighboring countries in Central and Eastern Europe, using the publicly available remotely sensed products, CORINE land cover and GHS-POP. The paper introduces a map modelling process that starts with data transformation through GIS analyses and results in the geovisualisation of the LCRPGR (land use efficiency), the PGR (population growth rate), and the LCR (land use rate). We investigated the spatial patterns of the index values by utilizing hotspot analyses, autocorrelations, and outlier analyses. The results show how the indicators’ values were concentrated in both countries; the average value of SDG 11.3. 1, from 2000 to 2018 in Poland amounted to 0.115 and, in Lithuania, to −0.054. The average population growth ratio (PGR) in Poland equaled 0.0132, and in Lithuania, it was −0.0067, while the average land consumption ratios (LCRs) were 0.0462 and 0.0067, respectively. Areas with an increase in built-up areas were concentrated mainly on the outskirts of large cities, whereas outliers of the LCRPGR index were mainly caused by the uncertainty of the source data and the way the indicator is interpreted.

Multitemporal Change Detection Analysis in an Urbanized Environment Based upon Sentinel-1 Data

The German Ruhr area is a highly condensed urban area that experienced a tremendous structural change over recent decades with the replacement of the coal and steel industries by other sectors. Consequently, a lot of major land cover changes happened. To retrospectively quantify such land cover changes, this study analysed synthetic aperture radar images of the Sentinel-1 satellites by applying the Google Earth Engine. Three satellite images are analysed by the multitemporal difference-adjusted dispersion threshold approach to capture land cover changes such as demolished buildings and new buildings by applying a threshold. This approach uses synthetic aperture radar data that are rarely considered in previously existing land cover change services. Urbanization or urban sprawl leads to changes in the urban form globally. These can be caused, for example, by migration or regionally by structural change, etc., such as in the study area presented here. The results are validated with reference data sets, which are publicly available nationally (e.g., house contour lines, normalized digital terrain model, digital orthophotos) or which are publicly available globally like the Global Urban Footprint and the World Settlement Footprint. Based on this, land cover changes could be identified for 21 locations within the study area of the city of Bochum.

Towards an Improved Large-Scale Gridded Population Dataset: A Pan-European Study on the Integration of 3D Settlement Data into Population Modelling

Large-scale gridded population datasets available at the global or continental scale have become an important source of information in applications related to sustainable development. In recent years, the emergence of new population models has leveraged the inclusion of more accurate and spatially detailed proxy layers describing the built-up environment (e.g., built-area and building footprint datasets), enhancing the quality, accuracy and spatial resolution of existing products. However, due to the consistent lack of vertical and functional information on the built-up environment, large-scale gridded population datasets that rely on existing built-up land proxies still report large errors of under- and overestimation, especially in areas with predominantly high-rise buildings or industrial/commercial areas, respectively. This research investigates, for the first time, the potential contributions of the new World Settlement Footprint—3D (WSF3D) dataset in the field of large-scale population modelling. First, we combined a Random Forest classifier with spatial metrics derived from the WSF3D to predict the industrial versus non-industrial use of settlement pixels at the Pan-European scale. We then examined the effects of including volume and settlement use information into frameworks of dasymetric population modelling. We found that the proposed classification method can predict industrial and non-industrial areas with overall accuracies and a kappa-coefficient of ~84% and 0.68, respectively. Additionally, we found that both, integrating volume and settlement use information considerably increased the accuracy of population estimates between 10% and 30% over commonly employed models (e.g., based on a binary settlement mask as input), mainly by eliminating systematic large overestimations in industrial/commercial areas. While the proposed method shows strong promise for overcoming some of the main limitations in large-scale population modelling, future research should focus on improving the quality of the WFS3D dataset and the classification method alike, to avoid the false detection of built-up settlements and to reduce misclassification errors of industrial and high-rise buildings.

Gridded population mapping for Germany based on building density, height and type from Earth Observation data using census disaggregation and bottom-up estimates

Gridded population data is widely used to map fine scale population patterns and dynamics to understand associated human-environmental processes for global change research, disaster risk assessment and other domains. This study mapped gridded population across Germany using weighting layers from building density, building height (both from previous studies) and building type datasets, all created from freely available, temporally and globally consistent Copernicus Sentinel-1 and Sentinel-2 data. We first produced and validated a nation-wide dataset of predominant residential and non-residential building types. We then examined the impact of different weighting layers from density, type and height on top-down dasymetric mapping quality across scales. We finally performed a nation-wide bottom-up population estimate based on the three datasets. We found that integrating building types into dasymetric mapping is helpful at fine scale, as population is not redistributed to non-residential areas. Building density improved the overall quality of population estimates at all scales compared to using a binary building layer. Most importantly, we found that the combined use of density and height, i.e. volume, considerably increased mapping quality in general and with regard to regional discrepancy by largely eliminating systematic underestimation in dense agglomerations and overestimation in rural areas. We also found that building density, type and volume, together with living floor area per capita, are suitable to produce accurate large-area bottom-up population estimates.

High-Resolution Gridded Population Datasets: Exploring the Capabilities of the World Settlement Footprint 2019 Imperviousness Layer for the African Continent

The field of human population mapping is constantly evolving, leveraging the increasing availability of high-resolution satellite imagery and the advancements in the field of machine learning. In recent years, the emergence of global built-area datasets that accurately describe the extent, location, and characteristics of human settlements has facilitated the production of new population grids, with improved quality, accuracy, and spatial resolution. In this research, we explore the capabilities of the novel World Settlement Footprint 2019 Imperviousness layer (WSF2019-Imp), as a single proxy in the production of a new high-resolution population distribution dataset for all of Africa—the WSF2019-Population dataset (WSF2019-Pop). Results of a comprehensive qualitative and quantitative assessment indicate that the WSF2019-Imp layer has the potential to overcome the complexities and limitations of top-down binary and multi-layer approaches of large-scale population mapping, by delivering a weighting framework which is spatially consistent and free of applicability restrictions. The increased thematic detail and spatial resolution (~10 m at the Equator) of the WSF2019-Imp layer improve the spatial distribution of populations at local scales, where fully built-up settlement pixels are clearly differentiated from settlement pixels that share a proportion of their area with green spaces, such as parks or gardens. Overall, eighty percent of the African countries reported estimation accuracies with percentage mean absolute errors between ~15% and ~32%, and 50% of the validation units in more than half of the countries reported relative errors below 20%. Here, the remaining lack of information on the vertical dimension and the functional characterisation of the built-up environment are still remaining limitations affecting the quality and accuracy of the final population datasets.

Towards a Large-Scale 3D Modeling of the Built Environment—Joint Analysis of TanDEM-X, Sentinel-2 and Open Street Map Data

Continental to global scale mapping of the human settlement extent based on earth observation satellite data has made considerable progress. Nevertheless, the current approaches only provide a two-dimensional representation of the built environment. Therewith, a full characterization is restricted in terms of the urban morphology and built-up density, which can only be gained by a detailed examination of the vertical settlement extent. This paper introduces a methodology for the extraction of three-dimensional (3D) information on human settlements by analyzing the digital elevation and radar intensity data collected by the German TanDEM-X satellite mission in combination with multispectral Sentinel-2 imagery and data from the Open Street Map initiative and the Global Urban Footprint human settlement mask. The first module of the underlying processor generates a normalized digital surface model from the TanDEM-X digital elevation model for all regions marked as a built-up area by the Global Urban Footprint. The second module generates a building mask based on a joint processing of Open Street Map, TanDEM-X/TerraSAR-X radar images, the calculated normalized digital surface model and Sentinel-2 imagery. Finally, a third module allocates the local relative heights of the normalized digital surface model to the building structures provided by the building mask. The outcome of the procedure is a 3D map of the built environment showing the estimated local height for all identified vertical building structures at 12 m spatial resolution. The results of a first validation campaign based on reference data collected for the seven cities of Amsterdam (NL), Indianapolis (US), Kigali (RW), Munich (DE), New York (US), Vienna (AT), and Washington (US) indicate the potential of the proposed methodology to accurately estimate the distribution of building heights within the built-up area.