Forrest R. Stevens

Dynamic population mapping using mobile phone data

Significance Knowing where people are is critical for accurate impact assessments and intervention planning, particularly those focused on population health, food security, climate change, conflicts, and natural disasters. This study demonstrates how data collected by mobile phone network operators can cost-effectively provide accurate and detailed maps of population distribution over national scales and any time period while guaranteeing phone users’ privacy. The methods outlined may be applied to estimate human population densities in low-income countries where data on population distributions may be scarce, outdated, and unreliable, or to estimate temporal variations in population density. The work highlights how facilitating access to anonymized mobile phone data might enable fast and cheap production of population maps in emergency and data-scarce situations. During the past few decades, technologies such as remote sensing, geographical information systems, and global positioning systems have transformed the way the distribution of human population is studied and modeled in space and time. However, the mapping of populations remains constrained by the logistics of censuses and surveys. Consequently, spatially detailed changes across scales of days, weeks, or months, or even year to year, are difficult to assess and limit the application of human population maps in situations in which timely information is required, such as disasters, conflicts, or epidemics. Mobile phones (MPs) now have an extremely high penetration rate across the globe, and analyzing the spatiotemporal distribution of MP calls geolocated to the tower level may overcome many limitations of census-based approaches, provided that the use of MP data is properly assessed and calibrated. Using datasets of more than 1 billion MP call records from Portugal and France, we show how spatially and temporarily explicit estimations of population densities can be produced at national scales, and how these estimates compare with outputs produced using alternative human population mapping methods. We also demonstrate how maps of human population changes can be produced over multiple timescales while preserving the anonymity of MP users. With similar data being collected every day by MP network providers across the world, the prospect of being able to map contemporary and changing human population distributions over relatively short intervals exists, paving the way for new applications and a near real-time understanding of patterns and processes in human geography.

A new urban landscape in East–Southeast Asia, 2000–2010

East–Southeast Asia is currently one of the fastest urbanizing regions in the world, with countries such as China climbing from 20 to 50 percent urbanized in just a few decades. By 2050, these countries are projected to add 1 billion people, with 90 percent of that growth occurring in cities. This population shift parallels an equally astounding amount of built-up land expansion. However, spatially-and temporally detailed information on regional-scale changes in urban land or population distribution do not exist; previous efforts have been either sample-based, focused on one country, or drawn conclusions from datasets with substantial temporal/spatial mismatch and variability in urban definitions. Using consistent methodology, satellite imagery and census data for greater than1000 agglomerations in the East–Southeast Asian region, the authors show that urban land increased to greater than 22 percent between 2000 and 2010 (from 155 000 to 189 000 square kilometers), an amount equivalent to the area of Taiwan, while urban populations climbed greater than 31 percent (from 738 to 969 million). Although urban land expanded at unprecedented rates, urban populations grew more rapidly, resulting in increasing densities for the majority of urban agglomerations, including those in both more developed (Japan, South Korea) and industrializing nations (China, Vietnam, Indonesia). This result contrasts previous sample-based studies, which conclude that cities are universally declining in density. The patterns and rates of change uncovered by these datasets provide a unique record of the massive urban transition currently underway in East–Southeast Asia that is impacting local-regional climate, pollution levels, water quality and availability, arable land, as well as the livelihoods and vulnerability of populations in the region.

Disaggregating Census Data for Population Mapping Using Random Forests with Remotely-Sensed and Ancillary Data

High resolution, contemporary data on human population distributions are vital for measuring impacts of population growth, monitoring human-environment interactions and for planning and policy development. Many methods are used to disaggregate census data and predict population densities for finer scale, gridded population data sets. We present a new semi-automated dasymetric modeling approach that incorporates detailed census and ancillary data in a flexible, “Random Forest” estimation technique. We outline the combination of widely available, remotely-sensed and geospatial data that contribute to the modeled dasymetric weights and then use the Random Forest model to generate a gridded prediction of population density at ~100 m spatial resolution. This prediction layer is then used as the weighting surface to perform dasymetric redistribution of the census counts at a country level. As a case study we compare the new algorithm and its products for three countries (Vietnam, Cambodia, and Kenya) with other common gridded population data production methodologies. We discuss the advantages of the new method and increases over the accuracy and flexibility of those previous approaches. Finally, we outline how this algorithm will be extended to provide freely-available gridded population data sets for Africa, Asia and Latin America.

High Resolution Population Distribution Maps for Southeast Asia in 2010 and 2015

Spatially accurate, contemporary data on human population distributions are vitally important to many applied and theoretical researchers. The Southeast Asia region has undergone rapid urbanization and population growth over the past decade, yet existing spatial population distribution datasets covering the region are based principally on population count data from censuses circa 2000, with often insufficient spatial resolution or input data to map settlements precisely. Here we outline approaches to construct a database of GIS-linked circa 2010 census data and methods used to construct fine-scale (∼100 meters spatial resolution) population distribution datasets for each country in the Southeast Asia region. Landsat-derived settlement maps and land cover information were combined with ancillary datasets on infrastructure to model population distributions for 2010 and 2015. These products were compared with those from two other methods used to construct commonly used global population datasets. Results indicate mapping accuracies are consistently higher when incorporating land cover and settlement information into the AsiaPop modelling process. Using existing data, it is possible to produce detailed, contemporary and easily updatable population distribution datasets for Southeast Asia. The 2010 and 2015 datasets produced are freely available as a product of the AsiaPop Project and can be downloaded from: www.asiapop.org.

Roads as drivers of change: Trajectories across the Tri-National Frontier in MAP, the Southwestern Amazon

Regional studies of land cover change are often limited by available data and in terms of comparability across regions, by the transferability of methods. This research addresses the role of roads and infrastructure improvements across a tri-national frontier region with similar climatic and biophysical conditions but very different trajectories of forest clearing. The standardization of methodologies and the extensive spatial and temporal framework of the analysis are exciting as they allow us to monitor a dynamic region with global significance as it enters an era of increased road connectivity and massive potential forest loss. Our study region is the “MAP” frontier, which covers Madre de Dios in Peru, Acre in Brazil, and Pando in Bolivia. This tri-national frontier is being integrated into the global economy via the paving of the Inter-Oceanic Highway which links the region to ports in the Atlantic and Pacific, constituting a major infrastructure change within just the last decade. Notably, there are differences in the extent of road paving among the three sides of the tri-national frontier, with paving complete in Acre, underway in Madre de Dios, and incipient in Pando. Through a multi-temporal analysis of land cover in the MAP region from 1986 to 2005, we found that rates of deforestation differ across the MAP frontier, with higher rates in Acre, followed by Madre de Dios and the lowest rates in Pando, although the dominant land cover across the region is still stable forest cover (89% overall). For all dates in the study period, deforestation rates drop with distance from major roads although the distance before this drop off appears to relate to development, with Acre influencing forests up to around 45 km out, Madre de Dios to about 18 km out and less of a discernable effect or distance value in Pando. As development occurs, the converted forest areas saturate close to roads, resulting in increasing rates of deforestation at further distances and patch consolidation of clearings over time. We can use this trend as a basis for future change predictions, with Acre providing a guide to likely future development for Madre de Dios, and in time potentially for Pando. Given the correspondence of road paving to deforestation, our findings imply that as road paving increases connectivity, flows of people and goods will accelerate across this landscape, increasing the likelihood of dramatic future changes on all sides of the tri‑national frontier.

Multitemporal settlement and population mapping from Landsat using Google Earth Engine

As countries become increasingly urbanized, understanding how urban areas are changing within the landscape becomes increasingly important. Urbanized areas are often the strongest indicators of human interaction with the environment, and understanding how urban areas develop through remotely sensed data allows for more sustainable practices. The Google Earth Engine (GEE) leverages cloud computing services to provide analysis capabilities on over 40 years of Landsat data. As a remote sensing platform, its ability to analyze global data rapidly lends itself to being an invaluable tool for studying the growth of urban areas. Here we present (i) An approach for the automated extraction of urban areas from Landsat imagery using GEE, validated using higher resolution images, (ii) a novel method of validation of the extracted urban extents using changes in the statistical performance of a high resolution population mapping method. Temporally distinct urban extractions were classified from the GEE catalog of Landsat 5 and 7 data over the Indonesian island of Java by using a Normalized Difference Spectral Vector (NDSV) method. Statistical evaluation of all of the tests was performed, and the value of population mapping methods in validating these urban extents was also examined. Results showed that the automated classification from GEE produced accurate urban extent maps, and that the integration of GEE-derived urban extents also improved the quality of the population mapping outputs.

High-resolution gridded population datasets for Latin America and the Caribbean in 2010, 2015, and 2020

The Latin America and the Caribbean region is one of the most urbanized regions in the world, with a total population of around 630 million that is expected to increase by 25% by 2050. In this context, detailed and contemporary datasets accurately describing the distribution of residential population in the region are required for measuring the impacts of population growth, monitoring changes, supporting environmental and health applications, and planning interventions. To support these needs, an open access archive of high-resolution gridded population datasets was created through disaggregation of the most recent official population count data available for 28 countries located in the region. These datasets are described here along with the approach and methods used to create and validate them. For each country, population distribution datasets, having a resolution of 3 arc seconds (approximately 100 m at the equator), were produced for the population count year, as well as for 2010, 2015, and 2020. All these products are available both through the WorldPop Project website and the WorldPop Dataverse Repository.

Amazon deforestation: Rates and patterns of land cover change and fragmentation in Pando, northern Bolivia, 1986 to 2005:

Much research has focused on deforestation in the Amazon, particularly with proximity to roads and population centers as proximate causes. This research presents the analysis of rates and patterns of land cover change in Pando, northern Bolivia, an area with most of its tropical humid forest still intact. Using a decision tree classifier, five forest/non-forest (FNF) classifications were created for 1986, 1991, 1996, 2000, and 2005 from 40 Landsat images that were preprocessed and mosaicked. FNF trajectory images were created for each date pair to indicate areas of stable forest and non-forest, and areas and rates of de/reforestation. Mean patch size, perimeter-area ratio, fractal dimension, and aggregation index metrics were calculated for the FNF trajectory images based on increasing buffer distances from road and along the main access road. In 2005, forest covered 95% of the area in Pando. Large areas of aggregated deforestation occur nearest the department capital of Cobija, along the border with Brazil, and about 50 km west and east of Cobija along the principal access road. Deforestation becomes patchier with increased distance from the population center and laterally from the road. Multiple non-linear relationships exist between the fragmentation metrics and distance from road. The results have implications for understanding and managing the spatial contiguity of these forests, which provide valuable ecological services as well as the livelihood base for many inhabitants.

Scaling categorical spatial data for earth systems models

Efforts to deduce the appropriate scales of ecosystem functions and how patterns change with scale have a long history in ecology and landscape ecology (Levin, 1992; O’Neill et al., 1996). Ecosystem function models are critical to predicting ecosystem responses to global change, but are limited by the technical challenges of model–data synthesis. Accurately relating phenomena across multiple scales is an important challenge in ecological modeling, as information is lost when converting between scales of analysis. Researchers must determine how much information is necessary to preserve the landscape signature of the ecological processes under study. Zhao & Liu (2014) sought to determine the appropriate spatial resolution for categorical land cover data to use in regional-scale models of carbon dynamics, and compared the use of two common categorical data resampling methods: majority (MR) and nearest neighbor (NNR). Their analysis of the NNR method showed a power-law relationship between study extent and grain, but results from MR method showed a different relationship, suggesting that the resampling method drove the results. Zhao & Liu (2014) concluded the NNR method to be superior and reported the MR approach produced ‘devastatingly deficient’ results. We discuss the lack of robustness of their power-law relationship by analyzing the configuration and composition of simulated landscapes subjected to different resampling methods. The authors stated that NNR is clearly preferential to the MR method because NNR preserves uncommon land cover types. They support their use of NNR by mis-citing Cain et al. (1997). Zhao & Liu (2014) state that the critical spatial resolution in scaling exercises follows a power-law function of the study region extent. We argue that the pattern of the landscape process to be modeled determines the results of the resampling method. We illustrate, using a simple simulated landscape, how the effect of resampling algorithm is related to the proportion of landscape within each land cover class and the spatial configuration (clumpiness)

Spatiotemporal patterns of population in mainland China, 1990 to 2010

According to UN forecasts, global population will increase to over 8 billion by 2025, with much of this anticipated population growth expected in urban areas. In China, the scale of urbanization has, and continues to be, unprecedented in terms of magnitude and rate of change. Since the late 1970s, the percentage of Chinese living in urban areas increased from ~18% to over 50%. To quantify these patterns spatially we use time-invariant or temporally-explicit data, including census data for 1990, 2000, and 2010 in an ensemble prediction model. Resulting multi-temporal, gridded population datasets are unique in terms of granularity and extent, providing fine-scale (~100 m) patterns of population distribution for mainland China. For consistency purposes, the Tibet Autonomous Region, Taiwan, and the islands in the South China Sea were excluded. The statistical model and considerations for temporally comparable maps are described, along with the resulting datasets. Final, mainland China population maps for 1990, 2000, and 2010 are freely available as products from the WorldPop Project website and the WorldPop Dataverse Repository.