Area and Population Estimates from Sentinel Data - Ghana and Mediterranean Egypt

Published: 25 January 2022| Version 1 | DOI: 10.17632/gf8v525tm6.1
Contributor:

Description

This dataset contains the area and population predictions made in the paper "A Method for Creating Globally Applicable Models for Population Estimates from Sentinel Data". The predictions are made using Sentinel 1 and 2 data and the newest available census data and population forecasts. The individual datasets are: ghana_area_float32.tif & egypt_area_float32.tif The predicted area of structures per pixel. Spatial Resolution: 10x10m Data format: float32 Valid for: 2021 Q3 Ghana, 2021 Q4 Egypt Nodata: -9999.0 ghana_area_uint8.tif & egypt_area_uint8.tif The predicted area of structures per pixel. Each pixel is 10x10m. The data format is uint8, rounded to the nearest integer from float32. Spatial Resolution: 10x10m Data format: uint8 Valid for: 2021 Q3 Ghana, 2021 Q4 Egypt Nodata: 255 ghana_population_unweighted.tif & egypt_population_unweighted.tif These are the unweighted population predictions for Ghana and Egypt. Unweighted means that daytime and nighttime population has not been taken into account, and each m2 of structure is weighted equally. Spatial Resolution: 10x10m Data format: float32 Valid for: 2021 Q3 Ghana, 2020 Q4 Egypt Nodata: -9999.0 ghana_population_unweighted_100m.tif & egypt_population_unweighted_100m.tif These are the same as ghana_population_unweighted.tif & egypt_population_unweighted.tif, but resampled to 100x100m using the SUM. Spatial Resolution: 100x100m Data format: float32 Valid for: 2021 Q3 Ghana, 2020 Q4 Egypt Nodata: -9999.0 ghana_area_residential.tif & ghana_area_non-residential.tif & ghana_area_residential_selforganised.tif This is the predicted area in m2 of the three classes: Residential, non-residential, and self-organised. It can be used to convert the area to a temporally adjusted population estimate as described in the paper. Spatial Resolution: 10x10m Data format: float32 Valid for: 2021 Q3 Ghana Nodata: -9999.9 ghana_m2_pr_person_conversion_layer.tif & egypt_m2_pr_person_conversion_layer.tif This is the predicted area pr person according to the most recent census and forecasts for each country. It can be used to convert from area to population. The Ghana numbers are based on ghana_area_float32 and the Ghana census 2021. Egypt's numbers are based on egypt_area_float32 and Egypt's 2017 census and UN population forecasts. The borders are smoothed with a 2km round kernel, but the sum is adjusted to remain the same. Spatial Resolution: 10x10m Data format: float32 Valid for: 2021 Q3 Ghana, 2020 Q4 Egypt Nodata: -9999.0

Files

Institutions

Aalborg Universitet

Categories

Remote Sensing, Population, Earth Observation, Deep Learning, Deep Transfer Learning

Licence