The role of data selection in mapping urban green and open spaces: a comparison across high and very-high resolution satellite imagery sources in two African cities
Bibliographic record
Abstract
Urban green and open spaces (UGOS) provide essential social, cultural, environmental, and economic benefits to a city; therefore, monitoring UGOS is critical for guiding management and strengthening urban resilience. Spatial analysis of Earth Observation data provides a practical means of evaluating UGOS, and with the availability of high and very-high spatial resolution (VHR) satellite imagery (≤10 m), UGOS can be accurately characterized across broad spatial and temporal scales. While VHR satellite imagery (≤3 m) can enable more refined characterizations of land cover (LC), its use may be constrained by high monetary costs, accessibility barriers, and reduced spatial and temporal coverage. This study investigates the implications of utilizing imagery sources of varying spatial resolution (≤10 m) and differing classification approaches—pixel-based versus object-based—on LC characterizations and subsequent UGOS spatial assessments in two urbanizing cities: Mekelle, Ethiopia and Polokwane, South Africa in 2020. LC classifications were derived from Sentinel-2 imagery (10 m), PlanetScope SuperDove imagery (3 m), and Maxar WorldView-3 multispectral (2 m) and pansharpened (0.5 m) imagery. Mapping accuracy and UGOS characteristics were evaluated for each city, including the composition of undeveloped versus developed land, tall vegetation cover, and LC within selected public spaces. Additionally, the share of streets and open space under Sustainable Development Goal Indicator 11.7.1 were assessed. WorldView-3 multispectral (2 m) LC maps consistently achieved the highest overall classification accuracies, at 92% in Mekelle and 86% in Polokwane, suggesting that spatial resolution alone does not guarantee higher mapping accuracy, and that spectral richness is an important characteristic for mapping complex vegetation. Although VHR imagery enhanced the detection of small and fragmented landscape features, such as trees, classification performance depended heavily on context, resolution, method, and image characteristics. Coarser imagery like Sentinel-2 proved to be practical for broader assessments (e.g., SDG 11.7.1) but based on our results, still may underrepresent total undeveloped space and fails to capture fine-scale spatial variation. The results revealed clearer spatial patterns and resolution-dependent trends in Mekelle, while findings in Polokwane were more variable, suggesting that local landscape structure and urban form may influence classification outcomes and UGOS metrics. Overall, this study highlights the importance of carefully selecting and interpreting Earth Observation imagery based on sensor characteristics, spatial and spectral resolution, classification method, acquisition timing, and local landscape context, especially when data options are limited.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.010 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.003 | 0.004 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".