Land Use Policy 133 (2023) 106842 Contents lists available at ScienceDirect Land Use Policy journal homepage: www.elsevier.com/locate/landusepol Planning for cooler cities in Ghana: Contribution of green infrastructure to urban heat mitigation in Kumasi Metropolis Isaac Sarfo a,m,1, Shuoben Bi b,c,*, Xiuhua Xu d, Emmanuel Yeboah e, Clement Kwang f, Michael Batame g, Foster Kofi Addai f, Umar Wakil Adamu f, Emmanuella Aboagye Appea f, Michael Atuahene Djan h, Henry Bortey Otchwemah i, Vanessa Elikem Kudoh f, Floribert Vuguziga j,n, Olumide Samuel Olowe k,2, John Ernest Koku l a College of Geography and Environmental Science, Henan University, 475004 Kaifeng, Henan, China b Research Base for Scientific Cognition and Protection of Culture Heritage, Nanjing University of Information Science & Technology, Nanjing, Jiangsu, China c School of Geographical Sciences, Nanjing University of Information Science and Technology, 210044 Nanjing, Jiangsu, China d Computer and Data Engineering College, NingboTECH University, Qianhu South Road, 315100 Ningbo, Zhejiang, China e School of Remote Sensing and Geomatics Engineering, Nanjing University of Information Science and Technology, 210044 Nanjing, Jiangsu, China f Department of Geography & Resource Development, University of Ghana, Legon, Ghana g Department of Natural Resources Management, University of Twente, Enschede, the Netherlands h Department of Geography, New Mexico State University, USA i Pheebes Consult Limited, Accra, Ghana j Key Laboratory of Meteorological Disaster of Ministry of Education (KLME), Collaborative Innovation of Meteorological Disasters (CIC-FEMD), Nanjing University of Information Science and Technology, Nanjing 210044 Jiangsu, China k Department of Animal Sciences, Purdue University, 270 S Russell Street, West Lafayette, IN, 47907, USA l Department of Social Sciences, Central University, Miotso-Tema, Ghana m Organization of African Academic Doctors (OAAD), Off Kamiti Road, P. O. Box 25305000100, Nairobi, Kenya n Rwanda Meteorology Agency (MeteoRwanda), Kigali, Rwanda A R T I C L E I N F O A B S T R A C T Keywords: This study investigates the spatial variability of some remote sensing indices representing built-up areas, vege- Garden city tation, bareness, and urban heat island (UHI), based on time-series Landsat TM/ETM+ and OLI/TIRS datasets Grand urban model archived for 1980–2020 period from the US Geological Survey’s website for Kumasi Metropolitan Area in Ghana. Land use Modules for Land Use Change Evaluation (MOLUSCE) and Cellular Automata Artificial Neural Network (CA- Urban heat island Urbanization ANN) algorithms and simulations in QGIS were used to predict future changes (2020–2050) for land-use systems in Kumasi. Findings revealed urbanization/built-up areas (+108.02%) contributed massively to the decline of forest areas (− 93.34%) and farmlands/shrubs (− 31.53%), thereby making Kumasi lose its once critical green position as the “Garden City of West Africa.” UHI moderately and strongly correlated positively against built-up (R2=0.78, p < 0.0001) and bareness (R2=0.96, p < 0.0001) indices, respectively. By contrast, UHI showed a statistically significant inverse relationship with the vegetative index (R2=0.97, p < 0.0001). Future land-use scenarios revealed more forests, waterbodies, and farmlands/shrubs will be lost, influencing urban tempera- ture and water supply. The multipurpose advantages of urban green space are ingrained in the grand urban model. Contextually, the Metropolis’s resilience has been hampered by inconsistency in the performance of institutional roles, competition for land ownership rights over green areas, and little investments or prioritization of green spaces. An integrated collaborative governance framework is proposed to unify actions, address power crisis and factors that influence governance of green infrastructure, UHI and land cover change. * Correspondence to: 210044 Nanjing, Jiangsu, China. E-mail addresses: sarfo.power@gmail.com (I. Sarfo), bishuoben@163.com (S. Bi), xiuhua_x@163.com (X. Xu), emmanuelyeboah@nuist.edu.cn (E. Yeboah), ckwang@ug.edu.gh (C. Kwang), michaelbatame8@gmail.com (M. Batame), fkaddai001@st.ug.edu.gh (F.K. Addai), umarwakil6030@gmail.com (U.W. Adamu), appeaella@gmail.com (E.A. Appea), michaelatuahenedjan@gmail.com (M.A. Djan), henrybortey@gmail.com (H.B. Otchwemah), kudohvanessa@gmail.com (V.E. Kudoh), vfloribert@nuist.edu.cn (F. Vuguziga), oloweolumides@gmail.com (O.S. Olowe), johnniekoku@gmail.com (J.E. Koku). 1 ORCID ID: https://orcid.org/0000-0002-6914-5764 2 ORCID ID:0000-0002-1677-492X https://doi.org/10.1016/j.landusepol.2023.106842 Available online 11 August 2023 0264-8377/© 2023 Elsevier Ltd. All rights reserved. I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 1. Background and related studies addresses UHI by promoting sustainable urban planning that prioritizes green spaces and vegetation. Trees, green walls and roofs, among other Land cover change (LCC), according to Stehfest et al. (2019) con- forms of vegetation have the potential to reduce urban temperature and notes a change in certain continuous characteristics of the land such as mitigate UHI effects. Green roofs and walls can also provide additional vegetation features, soil properties, and so on, whereas land-use change benefits such as energy conservation, pollution reduction, improve (LUC) encapsulates the transformation of a natural landscape into other public health and carbon sequestration. According to Teye (2018), forms. These alterations are accompanied by a host of regional/local and climate change threatens systems and processes in major cities. global implications, including variations in climatic conditions, biodi- In Ghana, akin to other nations in Africa, the capital cities are in most versity loss, health and safety, as well as ecosystem functions and end- instances challenged with the issue of rapid urban sprawl which con- points. Existing literature (Hong et al., 2022; Zhang et al., 2019) have tributes greatly to the intensity of UHI. Ghana’s overall population has reported the link between Land Use/Cover Change (LUCC) and climate grown almost five-fold in size between 1984 and 2021 (Ghana Statistical change. Growing concerns on prioritizing LUC amid sustainability Survey GSS, 2021). Again, the statistics of urban population over- concerns emanate from the said mesoscale phenomenon influencing shadowed the rural population growth by an annual percentage of 4.4. earth and atmospheric processes, sustainable development, food secu- Ghana’s urbanization rate according to the Ghana Statistical Service rity and public health (Zhang et al., 2019). Evidence of the dire conse- (GSS) rose from 30% to 57.35% (Ghana Statistical Survey GSS, 2021). quences associated with this phenomenon has driven stern initiatives The nation has initiated some socio-economic policies like the Ghana towards promoting sustainable practices, heat mitigation, land-use Vision 2020, Renewable Natural Resource Sector Strategy, Compre- management (LUM) and social innovations (SI). Notable among unde- hensive Development Framework (CDF), Ghana Poverty Reduction sired climate stressors linked to LUCC consist of urban heat island or Strategy (I & II), Millennium City Initiative, Savannah Accelerated amplification of land surface temperature, heat waves and certain forms Development Programme, Ghana Youth Employment and Entrepre- of diseases. Global warming and UHI effects are interconnected, but neurial Agency, National Urban Policy, Planting trees for Food and Jobs, ultimately different in horizon. The former is said to have emanated among others. Isioye et al. (2020) opined that in least developed nations, from severe city expansion during the industrial revolution, as observed the largest, medium and small-sized cities are the economic hubs. Early in the 20th century (Maxwell et al., 2018). The combination of the studies conducted in Kumasi Metropolitan Assembly (KMA) attempted foregoing factors led to the increasing range and ardency of UHI; a to quantify and spatially (Nero et al., 2017; Abass et al., 2018; Engel- phenomenon where obvious temperatures that are high are mostly hardt, 2012) present the distribution of UGS above and beneath the found in urban cities, while those in the rural areas are usually low due ground carbon pools, determine the extent of urban heat, effects of green to atmospheric and surface modifications associated with urbanization roofs on thermal discomfort in indoor spaces, and the perception of the (Long, 2022; Twumasi et al., 2021). Zhang et al. (2019) asserted that the populace in relation to their immediate settings and prevailing micro- effect of the preceding trigger of UHI threatens urban habitability to a climatic conditions (Koranteng et al., 2021). Studies conducted by great degree. They concluded that this phenomenon in the twenty-first Campion (2012) and Mensah et al. (2020) primarily delved into issues century is viewed as one of the major challenges for humankind. related to the sustenance of UGS, land use systems (Toure et al., 2020), Urban heat island (UHI) as an existential phenomenon of about 200 urbanisation, extreme events and wetlands. Twumasi et al. (2021) years has currently emerged, linked to habitability of urban cities. compared land surface temperature and lushness density of two major Presently, this phenomenon is considered as one of the primary issues cities in Ghana. In the same vein, Keith and Meerow (2022) proposed bedevilling the sustainability of cities. The event was identified over 200 some heat mitigation strategies that improves quality of life in cities. years ago in a study that describes the urban climate of London Abulibdeh (2021) posited that the degree of intensity and thermal bal- (Howard, 1818). Fundamentally, this issue has however become sig- ance of UHI remain a function of different components which comprise nificant due to the momentous expansion of urban cities and industri- urban design (location of open spaces and spacing between buildings), alization at a great scale. The event became noticeable during the past population growth and density, vertical and horizontal urban expansion, fifty years with devastating effects, considering a host of factors trig- topography, land use cover change attributes, and nature of building gering its size and intensity beside the common reasons like rapid materials (Estoque et al., 2017; Cui et al., 2016; Grover and Singh, population growth, industrialisation and automobiles, deforestation and 2015). extensiveness of agricultural activities, among other development pa- Kumasi Metropolis, earmarked among the major cities under the rameters. Data shows about 47% of Africa’s population live in urban Millennium City Initiative (MCI), has not received adequate attention on areas. United Nations Department of Economic and Social Affairs green infrastructure, UHI and its prognosis. The current study moves (UN-DESA, 2015) estimated the proportion of the world’s population beyond establishing a link between thermal comfort and vegetation, as living in urban areas amplified from 30% in 1950 to 54% in 2015. They well as identifying population growth and poverty as the sole drivers of further projected this figure to rise to 66% by 2050, citing rapid popu- LUCC and UHI in urban areas, as reported in numerous studies. In spite lation growth in Africa and Asia (UN-DESA, 2015). In this study, the of a surge in current literature and the guiding concepts highlighted in tenets of United Nations Millennium Sustainable Development Goals the previous paragraphs, particularly in developed countries, the task of (SDGs) 11 (targeted at renewing and planning cities/settlements to adhere to developing reliable predictions of future LUCC in KMA is further the principle of ensuring equitable distribution and access to basic services, to complicated by limited accurate baseline data on urban vegetative enhance human welfare and protect the environment), 13 (incorporates cover, urban hotspots and barren areas. The application value of this reactive and anticipatory measures into national policies that enhance study, specifically the land-use predictions will play a crucial role in adaptive capacity) and 15 (facilitate sustainable practices that protect and creating alternative futures and policy options, geared towards social, restore degraded lands) are drawn. Sustainable development remains a economic, institutional and environmental considerations. The outcome key challenge for the future of most cities, municipalities and metrop- of this study is vital in the sense that very few long-term analyses of olis. These guiding concepts are situated in the grand urban model changes in urban vegetation in African cities have been conducted. (GUM) which refers to the comprehensive planning and development of Again, examining some remotely sensed indices for deeper and clearer urban areas that are designed to be sustainable, eco-friendly and liveable insights for major cities remain an arduous challenge, coupled with (Hong et al., 2022; Çalışkan, 2012). One of the key concerns in the GUM allaying feasible and proactive strategies that addresses UHI effects and is the issue of UHI. UHI is a phenomenon where urban areas experience a urban green issues within the context of the grand urban model. significant increase in temperature due to rapid urbanization, and Standpoints reported in this study will be essential to urban commu- changes in land use patterns. UHIs are known to have significant impacts nities more widely amid climate concerns, both within and beyond the on public health, energy consumption, and the environment. GUM developing world. Therefore, we attempted to: 2 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 1. Analyze the spatial distribution of land use systems in Kumasi 2.2. Data acquisition and image classification Metropolis between 1980 and 2020. 2. Examine the spatial variability of urban vegetative cover, urban In this study, five Landsat images with a resolution of 30 m, archived hotspots, barren areas and UHI. for the given period (1980–2020), were acquired from the United States 3. Identify green and non-green hotspots that could be factored into Geological Survey’s (USGS) website (http://earthexplorer.usgs.gov/). sustainable urban planning. ArcGIS 10.8, ENVI 5.0 and 5.3 were used for the image pre-processing 4. Identify historical shifts in Normalized Difference Vegetative Index and enhancement. Table 1. (NDVI), Normalized Difference Built-up Index (NDBI) and Normal- ized Difference Bareness Index (NDBaI), and how they influence UHI. 2.2.1. Data processing 5. Create future LUCC scenarios (2020-2050) and propose policy op- Other image pre-processing procedures (Fig. 3) which were per- tions that address heat mitigation concerns. formed include image calibration, layer stacking and supervised classi- fication (Table 2). Bands 7, 4 and 2 were used for various classifications, 2. Study area and methods specifically for LANDSAT 4 (1980 s) and 5 TM (1990 s), as well as LANDSAT 7 ETM+ (2000 s and 2010), whilst LANDSAT 8 OLI/TIRS 2.1. Study setting (2020) used bands 5, 5 and 3 with 055/194 as row/path. We employed Landsat satellite imagery for this study due to its high level of stan- The Kumasi Metropolis (Fig. 1) lies 270 km north of Ghana’s capital. dardization, accessibility or provision of free datasets to individuals and With latitude 6◦40.664′N and longitude 1◦37.397′W. It occupies an area the global community, and its optimal ground resolution capacity to of about 214 km2. Kumasi Metropolis serves as the administrative cap- detect and monitor changes in landscapes and other ecosystems driven ital of the Ashanti region. by biophysical and other human-induced factors. The GSS (2021) estimates that KMA’s current population stands at 3, 490, 030 (Fig. 2). It links major areas to the northern and southern parts of Ghana, characterized by two main seasons (rain and dry seasons) with 2.3. Change detection analysis a transitional forest zone. Change detection analysis was run to ascertain the regularity of land use systems and its drivers in KMA, based on the following formulations: Fig. 1. Geographical location of Kumasi Metropolitan Area. 3 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 2. Total Population (millions) for Kumasi Metropolis (1984–2021). Table 1 Description of land cover types identified in KMA. Feature Definition Forests Closely interwoven trees and lush vegetation dominate these areas. It also includes all vegetative regions with no exposed soil. Built-up Urban, business, and industrial regions. This category also includes community green spaces, playing fields, and truck terminals. Bare land Bare sections of soil or rocks that haven’t been covered by greenery. In and around built-up regions, barren areas are noticeable. It constitutes terrains that have been cleared in preparation for redevelopment or cultivation. Farmlands and Widely distributed trees, hedges or bushes, secluded thickets, and non-tree crops. shrubs Waterbodies Rivers, lagoons, lakes, and other bodies of water are all part of this ecosystem. ( 2) LUCCChange in LUCC x Current year − LUCCPast year (1) 2.4.1. Coefficient of determinant = LUCCPast year Degree of proof or substantiation described as insignificant, rela- tively weak, or weak; least reasonable, or reasonable; and significant, or ( 2) LUCC − LUCC 2 % Change in LUCC x Current year Past year x 100% (2) substantial connection was determined through this research. R justi-= LUCCPast year fied the level of variation in a model. Thus, a model’s capacity to forecast [ ( ) ] or justify an event in a logistic regression scenario. Maximum R 2 score LUCCCurrent year − LUCCRateof change inLUCCperyear Past year x100% mean the deviation in an outcome may be clarified by estimation, using = LUCCPast year the specified parameters. R2 solely delves into magnitude of the asso- ÷40years ciation, and not ascertaining one variable as the cause of the other. 1 ≤ R2 (3) ≥0.8 means substantial or significant association between the given parameters. Where;. 0.7 ≤ R2 ≥0.5 indicates reasonable or least reasonable association. LUCCCurrent year denotes the final year under study, thus, the year 0.49 ≤ R 2 ≥0.2 indicates weak or relatively weak association. 2020 within the context of the present study; LUCCPast year denotes the Less than 0.2 indicates insignificant or no association. initial year being studied, thus, the year 1980. The rate of change in LUCC per year was estimated to ascertain the gains/expansion and 2.5. Land use prediction and validation losses/reduction in areas covered by the given land use classes over the past 40 years per the study duration. This provides detailed information Predictions for 2030, 2040, and 2050 were made using Modules for on how changes occurred annually to ascertain the major events, tran- Land Use Change Evaluation (MOLUSCE) in QGIS software version sitions and underlying mechanisms that drove such changes in order to 2.18.24. This plug-in uses Cellular Automata and Artificial Neural inform policies. Network (CA-ANN) techniques and simulations to forecast or make pre- dictions for KMA. The analysis primarily constitutes Evaluating Correla- tion (EC), area changes, Transition Potential Computation (TPC) 2.4. Spatial autocorrelation modelling and validation based on four (4) iterations. Digital Elevation Model (DEM) and a road raster georeferenced image of KMA were used as The value range for UHI, NDBI, NDVI and NDBaI was classified into reference data for the predictions. The main predictor variables used as twenty classes (Table 3) each using ArcMap 10.8. The relationship reference for future projections, based on the three (3) iterations consti- analysis was conducted to show whether there is a direct (positive) or tute built- environment (i.e., likelihood of change, density of developed inverse (negative) relationship between the given variables quantified in lands and crop land and transportation), socio-economy (i.e., population Table 3. The georeferenced or values generated for each of the under- density, number of households, urban population density, urbanisation studied variables were exported to Statistical Package for Social Sciences and industrialization), and natural environment (i.e., climatic variables- (SPSS) version 26 for further analysis. temperature, precipitation and moisture; ecology and topography). 4 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 3. Flow diagram of image-processing and post-classification for the study. In predicting LUCC in Kumasi Metropolis (Fig. 4), the model’s first 3. Results stage integrates LULC maps for the base periods (2010 and 2020) for predicate 2030, 2040 and 2050. Inputs are incorporated in the model to 3.1. KMA’s change detection analysis generate LCC map, from which the changing trends for the study setting between 2010 and 2020 are generated. These parameters constitute Statistics (Tables 4–5) presented indicate land cover conversion for DEM, distance from roads, and built-up density. Using Pearson’s cor- the five classes during the last four decades. Built-up and barren areas relation, Crammer’s coefficient and Joint Information uncertainty, the were the main land features that increased dramatically over the study link between geographic variables between the two raster images, period. Conversely, farmlands and shrubs, forests and waterbodies employed for this relationship is examined. Various classifications for experienced massive reduction in area coverage over the same period KMA along with the region’s dynamics are computed between the base (Fig. 5). period (2010) and the final year (2020). The TPC plug-in applies some machine learning algorithms like artificial neural networks (ANN), lo- 3.2. Urban Heat Island (UHI) analysis gistic regression (LC), weights of evidence (WoE), and some multi- criteria evaluation (MCE) to create TPC maps. These techniques or The mean temperature for the given study period (1980–2020) procedures draw on geographic information and LULC data to construct ranges between 31.62 ◦C to 20.79 ◦C. An upswing in UHI in Kumasi and calibrate changes in the region. ANN was applied to make pre- Metropolis as presented in Fig. 6 is primarily due to anthropogenic in- dictions for 2030, 2040 and 2050. A confusion matrix that merges and fluence, which is driven largely by population growth and distribution improves the user and producer accuracy assessments were used to (increasing settlements due to high birth/fertility rates and migration, gauge accuracy rates. urban sprawl, and so on), infrastructure development, and conversion of forest resources into other land cover forms. Tables 4–5 and Fig. 5 show 2.5.1. Accuracy assessment a remarkable rise in barren areas indicating the natural vegetation or In order to assess the accuracy of each study period, ground truth forests have been altered, with these areas cleared for development sample points were generated using ENVI and ArcGIS software. These purposes. More specifically, between 2010 and 2020, there was a points were overlaid on Google Earth Pro for verification. Twenty considerable rise (i.e., by 3.1 ◦C in maximum temperatures) in urban sample points were generated from each class in the classified images for heat, which could be attributed to the alarming rate of forest and land the accuracy assessment (see Fig.A.1). The expression or confusion degradation (declension rate of 2.3% between 1980–2020), and esca- matrix below was used to estimate the accuracy assessment: lating settlements (+108.02%), resulting in a significant increase in bare land (+924%), which influenced rising temperatures (i.e., 9.01 ◦C and Accuracy Assessment(A.A) = [(ASP∕TSP) × 100 ] (12) 5.2 ◦C as differences in maximum and minimum temperatures over the Where: past four decades) (Fig. 6) in the KMA. UHI effects influence morbidity A.S.P = Number of sample points that accurately fall on each and mortality of several ecosystems. required feature (ASP=94). 5 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Table 2 Quantification of remote sensing variables. Indices Formulation Source NDVI Abu Bakar et al. (2016) (NIR − RED) NDVI = (4) (NIR + RED) where; NIR= Near-infrared, RED= Red visible bands NDBI ( ) Xu (2007) (SWIR − NIR) NDBI = (5) (SWIR + NIR) NDBaI ( ) Chen et al. (2006) (Band5 − Band6) NDBaI = (6) (Band5 + Band6) For Landsat 8 data: ( ) (Band6 − Band10) NDBaI = (7) (Band6 + Band10) UHI analysis Avdan and Jovanovska (2016). (LMAXλ − LMINλ)Lλ = x(DN − QCALMIN)+LMINQCALMAX QCALMIN λ (8) ( − ) ( ) Where Lλ is cell value as radiance in W/ M2 ∗ sr ∗ µm ; ( ) LMAXλ is the sensor spectral radiance (Lλ) scaled to (QCALMAX) in W/ M2 ∗ sr ∗ µm ; LMINλ is the sensor spectral ( ) radiance that is scaled to (QCALMIN) in [W/ M2 ∗ sr ∗ µm ]. (QCALMAX) is the maximum quantized calibrated pixel value to LMAXλ [D.N], (QCALMIN) is the minimum quantized calibrated pixel value corresponding to LMINλ [D.N]; and QCAL is the quantized calibrated pixel value [D.N]. The LMIN and LMAX are the spectral radiances for each band at D.N 1 and 255 for Landsat 7 ETM+ , 1 and 65535 for Landsat 8 OLI/TIRS. λ is the wavelength. -Conversion of Lλ to Kelvin (K) with emissivity value: K T 2= ( )K E (9) ln 1 ∗ + 1 Lλ Therefore, k1 and k2 become coefficients determined by effective wavelength of a satellite sensor. K BT 2= (10) ln[(K1∕Lλ) + 1 ] Since temperature is required in Degree Celsius (◦C) (Tc), results for various temperatures must be converted from (K) satellite brightness temperature (TB) to (◦C) (Tc). TC = TB − 273.15 (11) 3.3. Green space and urban hotspot analysis Table 3 Autocorrelation analysis for the given indices and UHI. Over the research period, ArcMap was used to reclassify and assign UHI NDBI NDBaI NDVI lushness and non-greenness symbologies to vegetative and non- vegetative regions in the Kumasi Metropolis. Urban green studied in 36.02 0.99 0.96 -0.77 35.79 0.89 0.88 -0.72 KMA, as shown in Fig. 7, shows a constant decline in greenery. The city 35.43 0.69 0.78 -0.67 that was once dubbed the "Garden City of West Africa" because of its 34.88 0.59 0.68 -0.63 beautiful view of people living in harmony with nature has lost its 34.52 0.49 0.58 -0.57 identity to socio-economic growth and development. A close observa- 33.73 0.39 0.48 -0.47 tion for the less density or concentration at the heart of KMA between 32.98 0.29 0.38 -0.39 31.73 0.19 0.28 -0.27 1990–2000 (Fig. 8) is attributed to the decongestion policy/exercise that 30.87 0.09 0.18 -0.02 took place for the effective implementation of Renewable Natural 29.77 0 0 0 Resource Sector Strategy between 1996 and 2000, redevelopment of the 28.52 -0.002 -0.09 0.09 city, and concerns over decline of KMA’s lushness due to high influx of 28.12 -0.009 -0.19 0.19 27.73 -0.013 -0.29 0.29 migrants from nearby or other regions in such of greener pastures, 27.34 -0.044 -0.39 0.39 among other institutional and policy-driven factors. 26.85 -0.059 -0.49 0.49 Urban hotspot analysis (Fig. 8) was also performed to spatially depict 25.47 -0.065 -0.59 0.59 areas and the magnitude of urbanization between 1980 and 2020. Based 23.77 -0.18 -0.69 0.69 on the evidence presented in Tables 5–6, as well as Fig. 8, settlements 21.72 -0.2 -0.79 0.79 19.33 -0.21 -0.89 0.89 within the Metropolis have amplified drastically. The quest of bridging 18.38 -0.22 -0.99 0.99 housing deficit gaps, enhancing capital expenditure, as well as indus- trialization substantially support the spatial results generated by this study. KMA is Ghana’s second most populated and largest city, and a 6 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 4. Evaluation procedure for land-use change predictions. Table 4 Area coverage for each class (km2) (1980–2050). Class/Period 1980 s 1990 s 2000 s 2010 2020 2030 2040 2050 Built-up 120.4 170.91 193.19 237.67 250.46 262.66 275.45 287.57 Farmlands & shrubs 30.1 65.68 19.58 20.35 20.61 19.33 12.52 5.13 Forests 146.32 50.11 74.12 36.68 9.75 5.42 4.21 2.57 Waterbodies 0.44 0.57 1.15 0.41 0.36 0.29 0.25 0.19 Bare land 1.74 11.73 10.96 3.89 17.82 11.3 6.57 3.54 * **Total area coverage (Km2) (Absolute) = 299 Table 5 KMA’s temporal variations of land cover changes (LCC) in windows (1980–2050) (%). 7 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 5. LUCC over the past four decades in KMA. commercial hub for production and business transactions (commerce A close observation of Fig. 2, considering the spike in KMA’s popu- and services) after Accra Metropolitan Area. Since the population of lation over the past few decades is in tandem with Abass (2018) KMA is expanding at an annual rate of 2.3% between 1984 and 2021 standpoints. In his study dubbed “Peri-urbanisation and loss of arable (Fig. 3), demand for housing, infrastructure, food and employment, land in Kumasi Metropolis in three decades”, he asserted that urban among others, continuously increases. These persistent expansionary population across the globe are expected to grow by more than two forces exerted on the natural environment to meet the growing demands thirds. According to the GSS (2021) report on “Population projections/ or needs of the populace is what has resulted in the massive decline of prospects”, the Ashanti region of Ghana which has Kumasi Metropolis as forest or natural vegetation over the years with undesired consequences its administrative capital projects the population of the region to hit 9, like UHI. 607,389 by 2050. The region currently has a population of 5, 924, 294, with 3.49 million residing in KMA. Similarly, a study conducted by PricewaterhouseCoopers (PricewaterhouseCoopers PwC, 2022), a 3.4. LUCC prediction for Kumasi Metropolis multinational accounting and auditing firm in the United Kingdom, stipulates that “the current urban-rural split in Ghana which is 54% in The distribution (Tables 5–6) shows built-up will increase at a rate of urban areas and 46% in rural areas is likely to expand in favour of urban 14.82% (with 0.5% increment each year), farmlands and shrubs will areas in the next three decades.” According to PwC (2022) and GSS decline at a rate of 75.11% (with 2.5% decrease each year), areas (2021) reports, Ghana’s population is expected to surpass 50 million by covered by waterbodies will decrease by 47.22% (with 1.6% declension 2050. Out of this total, 36 million are expected to live in urban areas, each year), bare land will expand by 80.14% (at 2.7% decreasing rate thus, 73% of the total population are projected to be living in urban each year) whilst forests over the study period (2020–2050) will decline areas by 2050. Results from the present study’s land use predictions by 73.64% (at a declension rate of 2.5% each year). Considering current (2020–2050), which anticipates a continuous increase in built-up aligns trends of immense changes in KMA, the projected statistics or scenarios with the reports and assertions of PwC (2022) and GSS (2021). (Fig. 9) could occur if strategic and results-oriented measures are not designed to avert or regulate current anthropogenic activities driving these changes. The highlighted drivers and trends without regulation defeat the purpose of developing a sustainable or a resilient city that enhances social well-being and quality of life. 8 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 6. UHI variations in KMA (1980–2020). 3.5. Assessment of the various indices in KMA terrain are examples of such regions. Yellow or light yellowish regions indicate less plant cover. 3.5.1. Vegetative-index Fig. 10 illustrates a remarkable decline in vegetative index over the 3.5.2. Built-up index study period. Forest regions have higher NDVI values due to enhanced There is substantial evidence of settlement growth over the study green biomass of trees and other plants. Forest and wildlife reserves/ period (Fig. 11). Magenta denotes areas where the built-up environment parks, closed (dense) and open canopies make up the greenish regions. has undergone significant positive change (presence). Natural vegeta- Reduction in NDVI based on study findings could be attributed to ur- tion (forests) and farmlands/shrubs, respectively, are depicted by dark banization (increase in human settlements, migration and so on), gov- brown and yellowish sections. Light green zones are locations with few ernment policies, programs and projects, initiated between 1980 s- human populations and vast stretches of undeveloped land. 1990 s (i.e., 1983 Economic Recovery and Stabilization Program; Vision 2020 (1996–2020); Ghana Poverty Reduction Strategy I and II 3.5.3. Bareness-index ((2003–2005, 2004–2007)/Ghana Shared Growth Development Agenda The significant variations in bareness index across the research (2015–2020); National Urban Policy (2013 till date); Comprehensive period are illustrated in Fig. 12. Regions with minimal bareness (less Development Framework (1999–2001); Renewable Natural Resource exposed patches) are represented by deep-blue areas, which are covered Sector Strategy (1996–2000), as well as infrastructure development. by natural flora. Reddish zones depict barren regions, left empty or Reddish regions indicate a significant negative change in greenery unoccupied in order to be developed for infrastructure or farming. Built- throughout the research period. Built-up environments and barren up regions are represented by yellow patches and light red/yellow areas. 9 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 7. Green and non-green space zones in KMA. 3.6. Relationship between the given study variables relationship between NDBI and UHI. Findings reported in this study is in tandem with the results mentioned by the studies above. The necessity of creating a relationship (Figs. 13(a)-13c) between Plotting UHI values against the NDVI depicted a statistically signif- UHI and the specified remote sensing indices is critical in acquiring a icant strong inverse relationship (R2=0.966, p < 0.0001). Here, thorough understanding of land use and climate research. Identifying R2= 0.9662 x100% = 93.3%. This means that NDVI accounts for 93.3% and tracking changes linked with historical land use system evolution of the variance in UHI. The study found robust evidence in relation to validates strategic and results-oriented actions focused at reversing the link between the two variables. The inverse relationship in Fig. 13(b) these unfavorable happenings. Global environmental change is the suggests as NDVI decreases, UHI increases. Geospatial analysis based on outcome of a combination of these changes at the local and regional LUCC and UHI proves drastic decline in forests, and a significant change levels throughout the world. in surface temperature (Fig. 6). Furthermore, unlike forests or thick Generally, UHI and NDBI as reported in most studies have a positive vegetation, which substantially influence microclimatic conditions, correlation. The degree of correlation or association varies spatially farmlands and shrubs, despite their greenness, have little impact on from one geographical location to the other. In the context of KMA, a surface temperature. statistically significant moderate positive correlation was identified UHI and NDBaI plots in Fig. 13 (c) presents a strong positive rela- (R2=0.78, p < 0.0001). Here, R2= 0.782 x100% = 60.84%. This in- tionship (R2=0.957, p < 0.0001). Here, R2= 0.9572 x100% = 91.59%. dicates 60.84% of the variation in UHI is explained by NDBI (Fig. 13a). NDBaI is responsible for 91.6% of the variance in UHI. Bareness index Here, the moderate positive correlation generated connotes the study increases with urban heat in KMA, according to representations (Figs. 6, identified enough evidence in relation to the link between the two 12 and 13c). According to Mensah (2015), green zones within the KMA variables. Contextually, geographic analysis revealed a significant and have been encroached or removed for socio-economic gains. consistent rise in built-up (Fig. 11) and UHI (Fig. 6), respectively. Considering the growth in human settlements among other socio- economic activities in the study area, previous studies (Buo et al., 2021; Qin et al., 2020; Amos-Abanyie, 2009) have reported a direct 10 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 8. Urban concentration hotspots in KMA. Table 6 Rate and magnitude of change (sq.km) in LCC of Kumasi Metropolis. 11 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 9. Predicted future LUCC scenarios for KMA (2020–2050). 4. Discussion monitoring of sustainable policies and projects, deforestation, attitudes of individuals and private developers, lifestyles and cultural issues (is- 4.1. Drivers and temporal variations of KMA’s LUCC sues linked to conflict of power and ownership of assets, due to the cultural setting or monarchical system in the region), and state driven We identified over six major factors (proximate/ underlying) that policies or governance systems tilted towards neo-liberalism and capi- influence LUCC in KMA based on reports from other studies, major talism allay some economic and socio-political factors driving these events and the spatial results generated. Taking into account the overall undesirable changes. These six key issues, based on previous studies conversions of LUCC systems (as shown in Table 5 and 6), built up (Frimpong and Molkenthin, 2021; Abass et al., 2018; Nero et al., 2017) expanded at a rate of 108.02% (with 2.7% increment each year), drove massive changes in land use systems in Kumasi. The Metropolis farmlands and shrubs fell at a rate of 31.53% (with 0.8% decrease each once known as the “Garden City of West Africa” has lost its once critical year), waterbodies reduced by 18.18% (with 0.5% declension each ‘green position’, mainly as a result of these direct or underlying factors. year), bare land increased by 924.14% (at a rate of 23.1% increment each year) whilst forests over the study period (1980–2020) declined by 4.2. Factors influencing the development of green spaces in Kumasi 93.34% (at a declension rate of 2.3% each year). Results presented in Metropolis Tables 4–6 show an area coverage (sq.km) for each class and evidence of considerable LUCC patterns in KMA between 1980–2020 and After exploring existing studies and policies, our study presents seven 2020–2050. The main land use features that increased progressively key factors impeding the development of green spaces in Kumasi. These over the study period were built-up and barren areas. Forests, water- foregoing barriers are situated in global discourse and the GUM. bodies and farmlands and shrubs experienced gradual decline over the given periods. a. State policies, governance systems and technological advancement: The Presently, the major drivers of LUCC in KMA, identified in this study government of Ghana over the past few decades have initiated moves beyond identifying population growth and distribution (caused several developmental policies and governance systems, aimed at by migration and high birth/fertility rates) and poverty alleviation as enhancing industrialization (transforming the economy from a raw the sole or critical drivers driving LUCC. Infrastructure development, to a highly industrialized/structured economy) like the one district weak institutional structures in the formulation, implementation and one factory policy initiative, bridging housing deficit gaps, 12 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 10. NDVI variations for KMA (1980–2020). capitalism and neo-liberalism governance systems initiated after the interference, nepotism and cronyism were highlighted as key bar- year 2000, minimum wage threshold and so on. Post-famine period riers that influenced stringent or stern enforcement of development in the 1980 s-1990 s drove the country to set-out an Economic Re- controls over UGS. covery and Stabilization Program in 1983 that ensured micro-and- c. Issue of land tenure systems or administration: Identifying the de- macroeconomic stability (Abbam et al., 2018; Teye, 2018; Aryee- partments or individuals who have absolute control over some tey and Kanbur, 2007). These policies welcome or boost foreign or designated areas for conservation or green space lands was a major private investments. However, private investors in most scenarios constraint. Hammond (2011) and Mensah (2015) presented three often focus on profit-oriented or natural resource sectors like the forms of land owners operating in the study domain. First delved into minerals, oil and gas, timber, fisheries and aquaculture that deteri- “vested lands” thus, lands owned by the government of Ghana; the orate the natural environment. These initiatives go a long way to second linked to “stool lands” (i.e., lands owned by the chiefs who are increase and prioritize some sectors over the others that end up in the custodians of the land), and third classified as “public lands” which trading of some conservative areas for economic gains. are acquired by the state for public usage like schools, landfill sites, b. Laxity in the implementation and monitoring of development controls: transportation networks, markets, open spaces and so on. Their study Development controls are instruments used for the planning and revealed the Lands Commission among other state agencies have the management of cities. They are tailored or designed to enhance the legislative backing to manage lands in the metropolis. However, populace’s quality of life. In the context of KMA, Mensah (2015) authorities from this unit opined these only exist on paper and not in reported this as a predominant factor that hindered the development, reality as their roles are being impeded by the traditional authorities and affected the quality of green spaces. He further stated areas who in some sections of the constitution or local acts have the power demarcated on the city’s layout as natural reserves, parks and wet- to exercise some jurisdiction. The traditional authorities, in most lands were either non-existent or heavily encroached. In most instances, sell out green space lands to private developers without developing countries, public parks are found to exist on paper, but their consent. The cultural set-up of Kumasi (considering its are found to be non-existent in reality. Likewise, political monarchical system) which engulfs the cultural perspective of urban 13 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 11. NDBI variations for KMA (1980–2020). governance, is key in the management and planning of such and shrubs meant to beautify the city by the public through using facilities. such spots as walkways or for trading, coupled with freely leaving d. Urbanization: Based on population and geospatial statistics presented their animals to graze on green spaces. The case of Kumasi could be in Figs. 2 and 8, increasing settlements and infrastructure develop- likened to that of Nairobi city in Kenya, as reported by Makworo and ment have resulted in massive encroachment of several green spaces. Mireri (2011). Likewise, the city’s geographical location (Fig. 1) is equally advan- f. Low prioritization and investment in green spaces: Results presented in tageous, and a major contributory factor. KMA remains a hub for Fig. 7 show alteration and dynamic ebb in urban greening. A city that several business transactions and services, hence, attracting several was once tagged as the “Garden City of West Africa” has lost its unique immigrants from the northern and southern zones of Ghana. status to socio-economic development. Steady decline of green e. General public’s attitude: Development of several parks or green spaces as depicted in Figs. 6 and 9 show UGS development have not spaces in developed countries have instituted collaborative efforts or been prioritized or had limited budgetary allocations due to its inputs from the populace. Sensitizing the public and other relevant relevance to the city and country, not being realized. stakeholders on the importance of green spaces have fostered some g. Power relations, over lapse and poor coordination among state agencies sense of ownership and change in attitudes towards green spaces, over green spaces: Effective coordination is an essential tool in the limiting encroachment and pollution of these areas. Mensah (2015) management and planning of UGS, coupled with healthy and posi- reported in the KMA, authorities expressed concerns over the pub- tive institutional alliance as observed in Hannover (Germany), Zur- lic’s attitude towards the conservation of UGS. Destruction of lawns ich (Switzerland) among other developed cities across the globe as 14 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 12. NDBaI variations for KMA (1980–2020). reported by Carmona et al. (2004). Mensah (2015) reported officials green growth in our development agenda at all levels to attain the in state agencies that were relevant to the scope of this study rated multifunctional benefits of green spaces. KMA’s available green the level of coordination between them as poor. He ascribed these to space currently stands at 4 m2 per capita, which is below the World undefined and incoherence of responsibilities, limited recognition Health Organisation’s (WHO) standard threshold of 9 m2. and over lapse in the execution of roles. This issue has led to desta- 2. State agencies, land use planners and estate developers among other bilization and halting of several green space projects. interested parties can organize themselves to work as groups through “an integrated collaborative governance framework” that unifies ac- 4.3. Addressing major constraints impeding the development of green tions, address power crisis and the driving factors identified. Again, spaces in KMA such partnerships could be key in developing heat mitigation stra- tegies in our quest to achieve green or blue footprints (i.e., devel- We sought to highlight some mechanisms that could be incorporated oping a green economy). Considering the KMA earmarked under the to address bottlenecks hindering the development and sustenance of Millennium City project or initiative and in line with the concept of UGS in KMA. These measures are developed based on policy recom- the GUM, KMA through social inclusion and innovation, stakeholder mendations, early studies and deductions from the outcome of the participation through principled engagement and shared motiva- spatial analysis. Here, it is worth noting that the proposed measures are tion/common goal, and involvement/consultation of all key actors related to the seven factors that impede the development and preser- including interested parties or international donors and scientific vation or sustenance of UGS in Kumasi Metropolis. communities, can facilitate the realization of low-carbon, inclusive- ness (i.e., in line with Local Agenda 21) and resource efficiency to 1. Prioritisation and provision of more green spaces backed by restore the area’s once critical green position and sustainability. budgetary allocations and political will, will enhance availability 3. Periodic maintenance and expansion of existing green facilities are and accessibility to these areas. This, in essence, would integrate highly recommended to meet the growing needs of the general public 15 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 13. (a) - (c): Correlation between the given indices and UHI. or visitors. Some funds generated from such facilities could be set- 7. Address issues of land tenure systems, co-management and coordi- aside for such purposes. The government could also intervene to nation among state agencies whose functions directly or indirectly assist through further investments where necessary. influence UGS. 4. Behavioural or mass education on the need to conserve green spaces 8. Enhance human resource-base of green spaces, coupled with stake- through public campaigns, research and media outlets. Fig. 14 shows holder participation and involvement at all levels in the decision- the status of some public green infrastructure engulfed with filth or making process. encroached by others for development and religious purposes. 4.4. Contribution of urban greening to heat mitigation in Kumasi 5. Public-Private Partnership Agreements (PPPA) could be strength- Metropolis ened between state agencies, private organisations and local com- munities to enhance maintenance and patronage of these areas Green spaces are placed among the indicators within the SDGs for (Fig. 15). Award schemes could be designed for green space units to accomplishment by 2030. Recent studies have reported the critical role, enhance good conditions and quality of services. For instance; The played by green spaces in the actualisation of sustainable cities (Abass UK and Germany have green flagship awards for well-conditioned et al., 2018; Nero et al., 2017). Green spaces protrude diverse benefits to facilities, which could be adopted by KMA. ecological systems and processes. Notable among these are the quality of life or social well-being, harmonious environmental exchanges and 6. Strengthen institutions with logistics, legislative frameworks, invigoration of economic growth. appropriate funds and human resources to monitor and punish en- The inverse relationship between NDVI and UHI as presented in croachers of green spaces. Stringent regulations are needed to ensure Fig. 13(b), coupled with earlier reports (Koranteng et al., 2021; Nero effective implementation of development controls imbibed in exist- et al., 2017) demonstrating the regulation of prevailing microclimatic ing plans and legal frameworks. Creating more green belts will serve conditions by green spaces present another advantageous perspective of as checks for people or private developers not to go against or prioritising the development of green spaces (i.e., tree and vegetative overextend boundaries. cover) in cities. UHI within KMA has increased substantially between 1980 and 2020, as a result of several conversions of land cover systems. 16 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Fig. 14. The state of some green infrastructure in Kumasi Metropolis. The given frames depict the current state of “Children’s Park at Amakom-Asafo.”, providing evidence of encroachment or misuse of the area for other purposes, limited investments and poor maintenance culture, leaving these areas in a deplorable state, and far from serving its intended purposes. Fig. 15. Image of Rattray Park, Kumasi -A state and a private owned green space infrastructure that serves the general public. 17 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Table 7 Green infrastructure options, their impact and priority locations in terms of heat mitigation in KMA. Type Description Impact on UHI mitigation Priority locations Green belts Medium to large uninterrupted areas of -They can significantly mitigate UHI effects in KMA by *High-density housing. open land around the city, or its reducing surface temperatures and cooling the surrounding *Hotspot zones with medium-to-high heat districts or neighbourhoods areas through increased vegetation, and intensity. evapotranspiration. The presence of green open spaces in *Dead-end streets/spots urban environments helps absorb heat, lower ambient *Areas with high ecological value and locations temperatures, and create conducive microclimatic near educational, recreational and commercial/ conditions, thus, counteracting UHI effects caused by the centres. density of buildings, roads, and impermeable surfaces in *Areas with poor air quality and limited Kumasi Metropolis. biodiversity. Urban parks and Large parks, pocket parks, -They provide cover, evaporative cooling, and minimizes *Areas with high population density and urban forests neighbourhood green space, botanical surface temperatures. They absorb and dissipate heat to temperatures. gardens, cultivated forests, etc. create thermal comfort for residents and other ecosystems. *Areas with poor air quality, limited biodiversity and communities with limited access parks. *Under-utilized areas or dead-end streets/spots *Areas with high ecological value and locations near educational, recreational and commercial/ centres. Urban agricultural Meadows, orchards, arable land, -They mitigate UHI effects by expanding green spaces, and *Rooftop farming systems grasslands, agroforestry, community enhancing evapotranspiration. Their presence somewhat *Community and rain gardens horticulture gardens, etc. reduces temperatures, cools surrounding environments and *Apartments and condominiums with balconies/ counteracts UHI effects induced by built-up surfaces in terraces. cities. *Academic institutions can incorporate urban agriculture into their systems to teach students about nutrition, environmental sustainability and agriculture. *Sections within public spaces/parks can be allotted for urban agriculture to facilitate strong community engagement. Small urban green Green canopies, green pavement, -They critically influence UHI effects by providing *Middle of road intersections or along the roads (i. systems vertical greening, green urban localized cooling effects, and improved air quality. e., traffic islands). furniture, etc. *Street corners, installation on building walls/ fences, in-between buildings, roofs, along sidewalks, parking lots, etc. *Along the road, pits, walkways, alleyways, bicycle lanes, public spaces, etc. *Underutilized areas -Living plants (i.e., plant resilient species)/facades with living materials such as panels containing growing plants. Building- Intensive and extensive green roofs, -They mitigate UHI effects by providing additional *The rooftop of buildings integrated green green facades, green balconies vegetative cover to minimize surface temperatures in *Building facades systems urban areas. These covers absorb and dissipates heat, *Rain gardens which cools down the surrounding or overall ambience *Permeable pavements based on their extent or coverage in cities. *Balconies/terraces of buildings *Natural playgrounds or zones -Outdoor learning spaces for environmental education -Living plants (i.e., plant resilient species)/facades with living materials such as panels containing growing plants. Green water Rain gardens, green filters, stormwater -They facilitate natural water infiltration and retention, *Residential areas, along streets, roadways, public management green streets/catchment and which aid in cooling down urban surfaces by reducing heat- spaces, etc. systems infiltration basins. absorbing impervious areas and increasing evaporation. *Create infiltration basins under stormwater green streets. *Constructed wetlands in open spaces or near waterbodies *Green roofs on buildings to harvest and retain rainwater Creation/sustenance of urban ponds/lakes to serve as natural storage ad treatment systems. *Designation of natural drainage channels, detention basins, buffer zones along rivers/streams *Implementation in educational institutions for training and promotion of sustainable practices. Green-blue Lakes, ponds, streams, rivers and small -They combine green spaces and water features to *Green streets, corridors, green-blue roofs, infrastructure watercourses, and their banks, positively impact or mitigate UHI effects. They enhance riverfront/ parks along riverbanks, etc. systems wetlands, etc. cooling by providing shades, and evaporative cooling to *Creation and sustenance of wetlands. enhance carbon sequestration, improve air quality and *Green-blue educational facilities/campuses, create more thermal comfort in urban environments. alleyways, parking lots. *Coastal green spaces, public playgrounds, urban resilient parks, etc. Green-grey Parking pavements, playgrounds, sport -They synergize green spaces with traditional built -Parking lots with permeable pavements, rain infrastructure facilities, street/roads greenery (e.g., materials or vegetation with hard surfaces (i.e., grey gardens, green roofs, detention basins, etc. systems tree alleys), railroads greenery, etc. infrastructure) to minimize the impact of UHI significantly (continued on next page) 18 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Table 7 (continued ) Type Description Impact on UHI mitigation Priority locations through shading, and surface albedo improvement that *Creation of wetland parks enhances the overall urban climate positively. -Green-grey institutional designs Semi-natural Cemeteries, abandoned areas, -They have considerable impact on mitigating UHI effects *Neighbourhoods/community centres infrastructure brownfields, etc. by enhancing air circulation, the Metropolis’s *Ecological restoration sites, riparian buffer zones, sustainability and reducing heat aggravation. open spaces, greenbelts, transportation and wildlife corridors. *Reclaimed or degraded lands *Academic/research institutional sites *Cemeteries and barren areas *Semi-natural roofs and facades Private green House gardens, green fences, ground -Initiating vegetation in private zones contribute some *Residential, commercial and industrial buildings infrastructure green walls, etc. level of cooling down effects (i.e., relatively small/limited *Mainstreaming green walls/vertical gardens on impact) in urban areas. However, collective presence the exterior of buildings. instead of individual impact of each private green facility -Permeable pavements using permeable materials has the tendency to reduce the overall heat buildup in for walkways, driveways and so on. KMA. *Building walls, fences, hedges or roofs with plant resilient/drought tolerant plant species. *Permaculture gardens on private properties. *Private wetlands, backyard structures like sheds, pergolas and greenhouses. The mean maximum temperature in KMA has risen at a rate of 0.035 ◦C (Nero, 2017), where 90% of total emissions were linked to CO2 and loss each year, which is higher than the global warming rate of 0.0150 ◦C. of avenue trees along the roads and Central Business District (CBD), (Amos-Abanyie, 2009). Given the substantial loss in forest acreage, this green areas could improve air quality and architectural beauty of green is consistent with the conclusions presented in this study. building designs (spaces, vertical and horizontal expansions) (Koranteng Major cities in developed countries like Atlanta in the USA, Hong et al., 2019; Koranteng et al., 2021). They have the potential to increase Kong and Singapore where structural change has occurred, draw carbon sequestration, thereby improving climatic conditions and air development strategists, planners, local officials, researchers, and other quality, while preserving tree and animal biodiversity. More so, reflec- interested parties to the table to brainstorm, examine and work towards tive and cooling pavements/roofs and facades as reported in other a desired goal (Aflaki et al., 2017; Wang et al., 2013). Action plans and studies could be integrated in the urban landscape of KMA, ventilation monitoring systems are designed by key stakeholders to develop sus- corridors, efficient equipment and appliances. Presently, other notable tainable strategies that improve the citizenry’s well-being and comfort heat mitigation strategies such as the development of wind paths (i.e., amid UHI concerns. Information presented in this research will serve as expansion of cities require novel engineering approaches, landscape impetus in this emerging body of literature that acknowledges the sig- planning, capital intensive and policy-driven options to create pathways nificant role of urban greening in mitigating UHI effects. It is meant to for winds that significantly minimize high temperatures in some specific provide the scientific basis for planners, policy-makers and developers to locations for KMA based on UHI analysis (Fig. 6)), improvement in road devise innovative means of addressing some local environmental issues surfaces and public means of water-sprinkling are practiced in Japan which underlie UHI, LUCC among other climatic stressors that could and China. Considering the asphaltic nature of roads in KMA, porous affect major, medium and small-sized cities in future, especially in concretes/permeable paving materials and systems can be used to developing nations. replace the asphalts or impermeable concretes which amplifies heat. The The morphology of green areas plays a critical role in the regulation water sprinkling effect during summer in some developed coun- of UHI as indicated in Fig. 13(b). Policy documentation, institutional tries/cities like Japan/Tokyo (Japan Meteorological Agency (Japan reports and systematic literature reviews based on studies (Twumasi Meteorological Agency JMA, 2018), could be adapted considering KMA et al., 2021; Koranteng et al., 2021; Buo et al., 2021; Mensah, 2015) characterized by tropical or warm conditions. People across different conducted in the region were subjected to content analysis. Spatial neighbourhoods can sprinkle water on the streets to regulate evapora- analysis proved urban green hotspots have been declining in KMA, tion or enhance cooling effects. Similarly, state and private agencies whilst a fundamental drift in urban density and UHI is observed. Stra- within the Metropolis can adopt novel means sprinkle water on town- tegies like (i) increasing tree and vegetative cover (Keith and Meerow, ship or major roads to enhance cooling effects. These strategies can 2022; Grover and Singh, 2015) (ii) creation of green roofs, green facades inform the Government of Ghana’s (GoG) decision and plan to reduce and architectural or building designs for the amplification of solar emission rates by 15 to 45%. reflectivity (Arellano et al., 2018) (iii) installation of reflective or cool Table 7 present some green infrastructure options, their impact and roofs and pavements (Qin, 2015; Akbari and Matthews, 2012) (iv) priority locations in terms of heat mitigation that could enhance urban Shades and ventilation corridors (Keith and Meerow, 2022) (v) devel- resilience of Kumasi Metropolis. opment of ecological and green, blue or grey economies and footprints The proposed green infrastructure options/facilities presented in (vi) use of energy efficient appliances and equipment (Abulibdeh, 2021) Table 7 could inform policies and create alternative futures which are and (vii) heat minimization from automobiles and air-conditioning area specific or defined to mitigate urban heat concerns in KMA. The (Keith and Meerow, 2022). These innovative and notable heat mitiga- proposed measures also go a long way to trickle down other socio- tion strategies require effective planning, community or public partic- cultural, economic and environmental benefits. It is worth noting that ipation/involvement in urban heat governance, coupled with scientific multiple options, regarding the proposed measures can be combined or analysis through the use of remote sensing and geospatial techniques to utilized to address heat concerns in KMA considering the complexity of inform decisions about urban climatic maps. Keith and Meerow (2022) UHI effects or degree of impacts trickled down by each of these and Arellano et al., 2018 have recognised vegetation as one of the key measures. elements of urban space with cooling effects. Given the amplification of vehicular greenhouse emissions in KMA 19 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 5. Conclusion Declaration of Competing Interest This study uses satellite imagery and current literature to show how The authors declare that they have no known competing financial urban green space may be used to address UHI effects in major cities, interests or personal relationships that could have appeared to influence within the context of the grand urban model. Due to the city’s unregu- the work reported in this paper. lated and unplanned expansion, some unfavourable changes in land use and urban systems have occurred in Ghana’s Ashanti region, notably Data Availability Kumasi. Several climatic stressors have been compounded by institu- tional ineptitude and the inability to adequately manage urban green Data will be made available on request. areas, as well as enforcing land use restrictions. Findings will make meaningful contributions on topics such as integrated collaborative Acknowledgement governance, development of sponge, low-carbon and garden cities. The study results can be summarized as follows: To the editors and the anonymous reviewers, we are most grateful for their useful and insightful comments. • Kumasi Metropolis has lost its once critical green position as the “Garden City of West Africa.” Appendix A. Supporting information • Urban heat statistically correlated positively with built-up and bareness indices. Supplementary data associated with this article can be found in the • Land use predictions (2020-2050) show thermal comfort and urban online version at doi:10.1016/j.landusepol.2023.106842. temperatures in Kumasi will be adversely impacted. • Proposed heat mitigation strategies can facilitate the realization of sponge, low-carbon and garden cities. References Abass, K., Adanu, S.K., Agyemang, S., 2018. Peri-urbanisation and loss of arable land in The geographical setting and evaluation method used in Kumasi Kumasi Metropolis in three decades: evidence from remote sensing image analysis. Metropolis could be applied in other major cities across Africa and Land Use Policy 72, 470–479. https://doi.org/10.1016/j.landusepol.2018.01.013. beyond. Our study toughens conceptual frameworks on the management Abbam, T., Johnson, F.A., Dash, J., Padmadas, S.S., 2018. Spatiotemporal variations in rainfall and temperature in Ghana over the twentieth century, 1900–2014. Earth of green spaces, urban heat mitigation and land use systems. Similarly, it Space Sci. 5, 120–132. https://doi.org/10.1002/2017EA000327. explicitly enriches basic datasets for multiple comparative case sce- Abu Bakar, S.B., Pradhan, B., Lay, U.S., Abdullahi, S., 2016. Spatial assessment of land narios, and influence policy directives about the nature and challenges surface temperature and land use/land cover in Langkawi Island. IOP Conf. Ser.: of green spaces in Ghana to inform the decisions of development prac- Earth Environ. Sci. 37. https://doi.org/10.1088/1755-1315/37/1/012064. Abulibdeh, A., 2021. Analysis of urban heat island characteristics and mitigation titioners. Health and safety problems, which are critical to UHI related strategies for eight arid and semi-arid gulf region cities. Environ. Earth Sci. 80 studies, might be further investigated in future works. Future research (2021), 259. https://doi.org/10.1007/s12665-021-09540-7. might focus on converting barren areas into green spaces, as well as Aflaki, A., Mirnezhad, M., Ghaffarianhoseini, A., Ghaffarianhoseini, A., Omrany, H., Wang, Z.H., Akbari, H., 2017. Urban heat island mitigation strategies: a state-of-the- incorporating cemeteries, which were not included in our study, into the art review on Kuala Lumpur, Singapore and Hong Kong. Cities 62, 131–145. https:// broader green infrastructure of metropolitan areas. An integrated doi.org/10.1016/j.cities.2016.09.003. collaborative governance framework is proposed to unify actions, power Akbari, H., Matthews, H.D., 2012. Global cooling updates: reflective roofs and pavements. Energy Build. 55, 2–6. crisis and address the factors that influence urban greening, UHI and Amos-Abanyie, S., 2009. Climate change and housing in Kumasi. In: Adarkwa, K.K. (Ed.), land cover change. Future of the tree: Towards growth and development of Kumasi”. University Printing Press, Kumasi, pp. 195–211. Arellano, B., Roca, J., Batlle, E., 2018. Green areas and urban heat island: combining Funding remote sensed data with ground-based observations. Remote Sens. Model. Ecosyst. Sustain. XV (Conf. Pap. ) 1–17. https://doi.org/10.1117/12.2320999. This work was supported by the National Natural Science Foundation Aryeetey, E., Kanbur, R., 2007. Economy of Ghana: Analytical perspectives on stability, growth and poverty. Boydell & Brewer. https://doi.org/10.7722/j.ctt81fmh. of China (No. 41971340 and No.41271410). Avdan, U., Jovanovska, G., 2016. Algorithm for automated mapping of land surface temperature using LANDSAT 8 satellite data. J. Sens. 1, 1–20. https://doi.org/ CRediT authorship contribution statement 10.1155/2016/1480307. Buo, I., Sagris, V., Burdun, I., Uuemaa, E., 2021. Estimating the expansion of urban areas and urban heat islands (UHI) in Ghana. Nat. Hazards 105 (2), 1299–1321. https:// Isaac Sarfo: Visualization, Methodology, Formal analysis, Data doi.org/10.1007/s11069-020-04355-4. curation, Conceptualization, Writing - original draft, Writing - review & Çalışkan, O., 2012. Grand urban rules. J. Urban Des. 17 (1), 160–162. https://doi.org/ editing. Bi Shuoben: Validation, Supervision, Resources, Funding 10.1080/13574809.2012.646152. Campion, B.B., 2012. Urban Wetland Ecology and Floods in Kumasi, Ghana. University of acquisition. Xiuhua Xu: Writing – review & editing, Supervision, Proj- Bremen. Doctoral thesis (Unpublished), Germany, pp. 1–217. ect administration, Data curation. Emmanuel Yeboah: Writing – review Carmona, M., Magalhaes, C.D., Blum, R., Hopkins, J. (2004). “Is the Grass Greener…? & editing, Validation, Software, Methodology, Formal analysis, Data Learning from international innovations in urban green space management”. London: CABE Space. curation. Clement Kwang: Writing – review & editing, Validation, Chen, X.L., Zhao, H.M., Li, P.X., Yin, Z.Y., 2006. Remote sensing image-based analysis of Methodology, Formal analysis. Michael Batame: Methodology, Data the relationship between urban heat island and land use/cover changes. Remote curation. Foster Kofi Addai: Validation, Formal analysis. Umar Wakil Sens. Environ., Therm. Remote Sens. Urban Areas 104, 133–146. https://doi.org/ 10.1016/j.rse.2005.11.016. Adamu: Writing – review & editing, Formal analysis. Emmanuella Cui, Y.P., Xu, X.L., Dong, J.W., Qin, Y.C., 2016. Influence of urbanization factors on Aboagye Appea: Writing – review & editing, Formal analysis. Michael surface urban heat island intensity: a comparison of countries at different Atuahene Djan: Writing – review & editing, Formal analysis. Henry developmental phases. Sustainability 8, 706. Engelhardt, F., 2012. Creating an environmental geographic information system for the Bortey Otchwemah: Writing – review & editing, Formal analysis. city of Kumasi, Ghana. Diva-Portal 1–5. Vanessa Elikem Kudoh: Writing – review & editing, Formal analysis. Estoque, R.C., Murayama, Y., Myint, S.W., 2017. Effects of landscape composition and Floribert Vuguziga: Writing – review & editing, Formal analysis. pattern on land surface temperature: An urban heat island study in the megacities of Southeast Asia. Sci. Total Environ. 577 (2017), 349–359. Olumide Samuel Olowe: Writing – review & editing, Formal analysis. Frimpong, B.F., Molkenthin, F., 2021. Tracking urban expansion using random forests for John Ernest Koku: Writing – review & editing, Formal analysis. the classification of Landsat imagery (1986–2015) and predicting urban/built-up areas for 2025: a study of the Kumasi metropolis, Ghana. Land 10 (1), 1–21. https:// doi.org/10.3390/land10010044. Ghana Statistical Survey (GSS) (2021). “Population and Housing Census report”. Government of Ghana, Kumasi. 20 I. Sarfo et al. L a n d U s e P o l i c y 133 (2023) 106842 Grover, A., Singh, R.B., 2015. Analysis of urban heat island (UHI) in relation to surface temperature of Kumasi. Cogent Environ. Sci. 6 (1) https://doi.org/10.1080/ normalized difference vegetation index (NDVI): a comparative study of Delhi and 23311843.2020.1787738. Mumbai. Environments 2015 (2), 125–138. Mensah, C.A. (2015). “Sustaining Urban Green Spaces in Africa: A case study of Kumasi Hammond, D.N.A., 2011. Harmonising land policy and the law for development in Metropolis, Ghana”. University of Birmingham, UK, Doctoral thesis (Unpublished Kumasi. In: Adarkwa, K.K. (Ed.), Future of the tree: Towards growth and thesis). development of Kumasi. University Printing Press, Kumasi, pp. 55–68. Nero, B.F., Callo-Concha, D., Anning, A., Denich, M., 2017. Urban green spaces enhance Hong, W., Wang, W., Guo, R., 2022. Policies for optimizing land-use layouts in highly climate change mitigation in cities of the global south: the case of Kumasi, Ghana. urbanized areas: an analysis framework based on construction land clearance. Procedia Eng. 198, 69–83. https://doi.org/10.1016/j.proeng.2017.07.074. Habitat Int. 130 (2022), 1–10. https://doi.org/10.1016/j.habitatint.2022.102697. PricewaterhouseCoopers (PwC) (2022). 2022 Global Annual Review: Ghana’s population Howard, L. (1818). “The climate of London: Deduced Meteorological observations at to surpass 50m by 2050. Accessed on July 21 2023 from 〈https://myjoyonline. different places in the neighbourhood of the metropolis”. Harvard University, 2, com/ghanas-population-to-surpass-50m-by-2050–36m-to-live-in-urban-areas-report Band 2 London. /#:~:text=Ghana%27s%20population%20is%20expected%20to,and%2037.83% Isioye, O.A., Ikwueze, H.U., Akomolafe, E.A., 2020. Urban Heat Island Effects and 20million%20in%202030〉. Thermal Comfort in Abuja Municipal Area Council of Nigeria. FUTY J. Environ. 14 Qin, Y., 2015. A review on the development of cool pavements to mitigate urban heat (2), 1–16. island effect. Renew. Sustain. Energy Rev. 52, 445–459. Japan Meteorological Agency (JMA). (2018). Japan for Sustainability. Tokyo Stehfest, E., van Zeist, W.J., Valin, H., et al., 2019. Key determinants of global land-use Metropolitan Environment Bureau. Available online: 〈https://resources.realestate.co projections. Nat. Commun. 10 (2019), 2166. https://doi.org/10.1038/s41467-019- .jp/living/urban-heat-island-effect-why-its-so-hot-in-tokyo-whats-being-done-abo 09945-w. ut-it/〉 (accessed on 21 July 2023). Teye, J., 2018. Urbanization and Migration in Africa. Univ. Ghana, U. Nations Hqrs. N. Y. Keith, L., Meerow, S., 2022. PAS report 600: planning for urban heat resilience. Am. 1-2 (November), 1–20. Plan. Assoc., Chic., USA 1–101. Toure, S.I., Stow, D.A., Clarke, K., Weeks, J., 2020. Patterns of land cover and land use Koranteng, C., Simons, B., Nyame-Tawiah, D., 2019. Green to grey: an urban heat change within the two major metropolitan areas of Ghana. Geocarto Int. 35 (2), assessment of Kumasi, Ghana. Int. J. Environ. Clim. Change 9 (12), 751–763. 209–223. https://doi.org/10.1080/10106049.2018.1516244. https://doi.org/10.9734/ijecc/2019/v9i1230155. Twumasi, Y.A., Merem, E.C., Namwamba, J.B., Mwakimi, O.S., Ayala-Silva, T., Koranteng, C., Simons, B., Nyame-Tawiah, D., 2021. Simulation-based analysis of the Frimpong, D.B., et al., 2021. Estimation of land surface temperature from landsat-8 effect of green roofs on thermal performance of buildings in a tropical landscape. OLI thermal infrared satellite data. A comparative analysis of two cities in Ghana. J. Innov. Sustain. RISUS 12 (1), 45–56. https://doi.org/10.23925/2179- Adv. Remote Sens. 10 (4), 131–149. https://doi.org/10.4236/ars.2021.104009. 3565.2021v12i1p45-56. United Nations, Department of Economic and Social Affairs, Population Division (UN- Long, H., 2022. Theorizing land use transitions: a human geography perspective. Habitat DESA), 2015. World Urban. Prospect.: 2014 Revis., (ST/ESA/Ser. A/366) 7–474. Int. 128 (2022), 1–11. https://doi.org/10.1016/j.habitatint.2022.102669. Wang, H., Shen, Q., Tang, B.S., Skitmore, M., 2013. An integrated approach to supporting Makworo, M., Mireri, C., 2011. Public open spaces in Nairobi City, Kenya, under threat. land-use decisions in site redevelopment for urban renewal in Hong Kong. Habitat J. Environ. Plan. Manag. 54 (8), 1107–1123. Int. 38, 70–80. Maxwell. K., Julius, S., Grambsch, A., Kosmal, A., Larson, L., Sonti, N. (2018). “Built Xu, H., 2007. Extraction of urban built-up land features from Landsat imagery using a environment, urban systems, and cities. In Impacts, Risks, and Adaptation in the thematic-oriented index combination technique. Photogramm. Eng. Remote Sens. United States: Fourth National Climate Assessment”. U.S. Global Change Research 73, 1381–1391. https://doi.org/10.14358/PERS.73.12.1381. Program, Washington, DC, 438–478. Zhang, Y.W., He, S., Gu, Z.L., Wei, N., Yu, C.W., Li, W., et al., 2019. Measurement, Mensah, C., Atayi, J., Kabo-Bah, A.T., Švik, M., Acheampong, D., Kyere-Boateng, R., normalization and mapping of urban-scale wind environment in Xi’an, China. Indoor et al., 2020. Impact of urban land cover change on the garden city status and land Built Environ. 28, 1171–1180. 21