Rapid urban growth in major cities of a country poses challenges for sustainable development. Particularly in Africa, the process of rapid urbanization is little understood and research is mostly limited to single cities. Thus, this study provides a comprehensive comparative analysis of the growth and spatial patterns of urban development in the three major cities of Ethiopia (Addis Ababa, Adama, and Hawassa) from 1987 to 2017. Also, the applicability of diffusion and coalescence theory on the evolution of these cities has been tested. Remote sensing and GIS technologies were combined with spatial metrics and morphological analysis was employed to undertake this study. The result revealed that all the studied cities experienced accelerated growth in the urbanized areas, but the cities with a larger initial urbanized size were associated with lower expansion rates. Differences in extent and direction of expansion in each city were mostly related to physical features, urban master plans, and policies, with an increase in the irregularity and dispersion of urban growth, representing strong evidence of urban sprawl. The spatiotemporal analysis confirmed that the urbanization processes of Addis Ababa and Adama were consistent and Hawassa city diverged from expectations based on diffusion and coalescence theory. In general, large cities with strong economic growth in a country fail to effectively control the scattered nature of urban growth, thus requiring aggressive policy intervention. The approach used in this study permits a deeper exploration of urban development patterns and the identification of priority areas for effective urban planning and management.

Fundamental ideas concerning urbanization are primarily based on studies performed in large cities. It is of interest to study whether or not similar phenomena take place in smaller cities. Small cities are an inherent component of urbanization, and in the future, the majority of globalization is expected to occur in small and mid-sized cities. Understanding the effects of small cities on landscape structures is, therefore, an essential component in planning city land expansion. Accordingly, this study focused on six towns of the Oromia Special Zone Surrounding Finfinnee, Ethiopia, which is broadly known to be experiencing dramatic growth. Time-series Landsat images from 1987 to 2019 with an integrated method, landscape metrics, and built-up density analysis were employed to characterize and compare the dynamics of landscape structures, urban expansion patterns, process, and overall growth status in the towns. The results highlight that all the towns experienced accelerated growth in the built-up areas and highly scattered nature in spatial growth. Landscape ecology analysis confirmed a highly fragmented urban landscape, a significant loss of natural land covers, and disconnected and complicated agro-vegetation patches in all towns, suggesting a lack of rigorous implementation of the master plan. Results also indicated that the Oromia Special Zone surrounding Finfinnee has failed to control urban sprawl to surrounding ecological sensitive areas. The study results, more broadly, highlight that the small cities would have a limited physical and demographic footprint and relatively less contribution to the national economic agglomeration; nonetheless, they can have a notable and important impact in terms of their ecological and environmental influence. Hence, the study suggests policies for monitoring such dynamics and protecting agro-environmental connectivity with particular focus on the small cities.

Urban green space (UGS) is considered a mitigative intervention for urban heat. While increasing the UGS coverage is expected to reduce the urban heat, studies on the effects of UGS configuration have produced inconsistent results. To investigate this inconsistency further, this study conducted a multi-spatial and multi-temporal resolution analysis in the Addis Ababa city metropolitan area for assessing the relationship between UGS patterns and land surface temperature (LST). Landsat images were used to generate land cover and LST maps. Regression models were developed to investigate whether controlling for the proportion of the green area (PGS), fragmentation, shape, complexity, and proximity distance can affect surface temperature. Results indicated that the UGS patches with aggregated, regular and simple shapes and connectivity throughout the urban landscape were more effective in decreasing the LST as compared to the fragmented and complicated spatial patterns. This finding highlighted that in addition to increasing the amount of UGS, optimizing the spatial structure of UGS, could be an effective and useful action to mitigate the urban heat island (UHI) impacts. Changing the spatial size had a significant influence on the interconnection between LST and UGS patterns as well. It also noted that the spatial arrangement of UGS was more sensitive to spatial scales than that of its composition. The relationship between the spatial configuration of UGS and LST could be changed when applying different statistical methods. This result underlined the importance of controlling the effects of the share of green spaces when calculating the impacts of the spatial configuration of UGS on LST. Furthermore, the study highlighted that applying different statistical approaches, spatial scale, and coverage of UGS can help determine the effectiveness of the association between LST and UGS patterns. These outcomes provided new insights regarding the inconsistent findings from earlier studies, which might be a result of the different approaches considered. Indeed, these findings are expected to be of help more broadly for city planning and urban heat mitigation.

Soil moisture (SM) is a critical factor influencing plant growth, agricultural yield, ecosystem functions, and water resource management. Existing soil moisture products, such as SMAP, ERA5, and ESA-CCI, provide daily soil moisture data. However, their coarse spatial resolution limits their application, necessitating downscaling for improved effectiveness. This study integrates multi-source remote sensing, reanalysis, and ground observation data through machine learning models to generate two-layer soil moisture data, aiming to improve the spatial resolution and accuracy of soil moisture data. First, the ESTARFM spatiotemporal fusion algorithm was applied to combine high-resolution MODIS data with long-term GLASS data, generating daily soil moisture driving variables (NDVI and LST) at a 1-kilometer resolution. Subsequently, the LightGBM downscaling algorithm was used to reduce the spatial resolution of GLEAM SSM and RZSM data from 0.25° to daily 1-kilometer resolution. Finally, a bias correction method based on convolutional neural networks (CNN) with transfer learning was employed for point-to-area fusion calibration to improve data accuracy. Experimental results show that the Pearson correlation coefficient (PCC) of SSM and RZSM data downscaled by LightGBM are 0.699 and 0.754, with root mean square error (RMSE) values of 0.053 m³/m³ and 0.055 m³/m³, respectively. After calibration with ground-based observations, the PCC of the soil moisture data ranges from 0.792 to 0.860, and the RMSE values range from 0.028 m³/m³ to 0.031 m³/m³, showing significant improvement in accuracy. Further spatiotemporal and comparative analyses confirm that the generated two-layer soil moisture data excels in capturing spatial and temporal dynamics. The study successfully generated high-precision, long-term time series soil moisture data for the Korean Peninsula, providing reliable support for agricultural drought monitoring and drought forecasting, and offering valuable references for soil moisture research in similar regions.