Carbon Storage and Land Use Dynamics in Ghanaian University Campuses: A Scenario-Based Assessment Using the InVEST Model

Abstract

University campuses in rapidly urbanizing regions face increasing pressure to balance infrastructure development with environmental sustainability, yet their carbon storage potential remains largely unexplored in sub-Saharan Africa. This study assessed land use changes, carbon storage dynamics, and economic valuation across three Ghanaian universities, University of Ghana (UG), Kwame Nkrumah University of Science and Technology (KNUST), and University of Cape Coast (UCC), from 2017 to 2023, and evaluated five future scenarios using the InVEST carbon model. Land use analysis employed ESRI 10 m annual land cover data, while carbon storage was estimated using regionally appropriate carbon pool values, and economic valuation applied Ghana’s social cost of carbon ($0.970/tCO₂). Historical analysis revealed substantial carbon losses: UG declined by 17.1% (19,695 Mg C), KNUST by 29.5% (20,063 Mg C), and UCC by 7.9% (3292 Mg C), due to tree cover conversion to built areas. Scenario modeling demonstrated that infrastructure-focused development would cause additional losses of 4211–6891 Mg C, while extensive tree expansion could increase storage by 1686–5227 Mg C. Economic analysis showed tree expansion generating positive net present values ($1612–$5070), while infrastructure development imposed costs (−$4028 to −$6684). These findings provide quantitative evidence for sustainable campus planning prioritizing carbon conservation in tropical institutional landscapes.

1. Introduction

Many terrestrial ecosystems serve as critical carbon sinks, with global forests and other vegetated landscapes absorbing approximately 3.2 ± 0.8 GtC annually from atmospheric CO₂ emissions over the last decade, and 3.8 ± 0.8 GtC in 2017, playing an essential role in global climate regulation [1,2]. These ecosystems sequester carbon through photosynthesis and store it in above-ground biomass, below-ground biomass, soils, and dead organic matter pools, creating long-term carbon reservoirs that help mitigate climate change impacts [3,4,5,6]. However, anthropogenic land use changes, particularly urbanization [7], agricultural expansion [8], and deforestation [9], severely compromise these ecosystems’ carbon sequestration capacity, with the Intergovernmental Panel on Climate Change (IPCC) estimating that Agriculture, Forestry, and Other Land Use (AFOLU) activities contribute approximately 12% (range 11–13%) of global anthropogenic greenhouse gas emissions [10]. The ecological consequences of land use change are especially severe in tropical regions, which host the planet’s richest biodiversity and largest carbon reserves. Between 2015 and 2020, approximately 17 million hectares of tropical forest were lost annually, resulting in substantial emissions from stored carbon and a permanent reduction in future carbon sequestration capacity [11,12,13,14].

West Africa exemplifies these regional challenges, experiencing some of the highest deforestation rates globally with approximately 3.9 million hectares of forest cover lost annually [15]. The region’s rapid urbanization, with urban populations growing at rates exceeding 4% annually, creates sprawling, unplanned development patterns that consume large areas of agricultural and forest land while transforming carbon sinks into carbon sources [16]. Ghana represents a particularly clear example of these dynamics, having lost nearly 25% of its tree cover from 2001 to 2023, largely due to urban expansion around major cities like Accra and Kumasi, agricultural development, and infrastructure projects [17]. Ghana’s forest ecosystems demonstrate significant carbon sequestration potential, particularly within urban green spaces and natural forest reserves. Studies in Kumasi metropolitan parks report average carbon stocks of 53.88 ± 11.70 MgC/ha, highlighting the role of urban biodiversity in sustaining ecosystem services [18]. Additionally, recent assessments of agroforestry systems in Ghana and Côte d’Ivoire suggest that increasing shade-tree cover could sequester up to 307 MtCO₂e, underscoring the untapped potential of Ghanaian landscapes to contribute meaningfully to climate mitigation efforts [19]. Studies across West African cities reveal significant urban carbon storage potential, with [20] recording 657.05 tonnes of total carbon storage across 836 tree stems in urban Zaria, Nigeria, and [21] finding mean carbon stocks of 31.63–58.30 t/ha in Niamey and Maradi, Niger. Within this broader context, urban ecosystems represent a particularly important but understudied component of carbon dynamics, with urban green spaces, including parks, gardens, street trees, and institutional landscapes, providing significant ecosystem services including carbon sequestration, air quality improvement, and temperature regulation [22].

University campuses represent a distinctive category of urban green space due to their large size, managed landscapes, mix of built and natural environments, and explicit institutional sustainability mandates that has received limited research attention despite their potentially significant role in urban carbon storage. These institutional landscapes often maintain substantial tree cover, botanical gardens, and integrated green spaces that may function differently from both natural forests and typical urban parks. Universities typically have longer planning horizons, greater institutional capacity for environmental stewardship, and explicit commitments to sustainability that could support enhanced carbon storage and sequestration. Moreover, university campuses often occupy large areas within rapidly urbanizing regions, making their land use decisions particularly consequential for regional carbon budgets. Recent studies have begun to explore this potential, with [23] assessing 17 campuses in Bangkok and finding average carbon stock densities of 46.77 tons/ha, while [24] emphasized the untapped potential of university-managed forests for climate mitigation and sustainability education.

The integrated methodology of analyzing land-use change, quantifying carbon storage with the InVEST model, and conducting economic valuation has been successfully applied across diverse global contexts, demonstrating its robustness for ecosystem service assessments. Several recent studies have applied integrated methodologies similar to this research, combining land use change analysis, carbon storage modeling using the InVEST model, and economic valuation to assess ecosystem services across diverse landscapes. Sharma et al. [25] conducted a spatio-temporal assessment of carbon storage in Noida, India, revealing a 67% increase in carbon storage between 2011 and 2019, followed by a projected decline by 2027 due to urban expansion. Rachid et al. [26] evaluated urban greening scenarios in Nador, Morocco, and found carbon storage increases of up to 70.3%, with economic benefits ranging from $2.01 to $3.6 million depending on the greening strategy. Kohestani et al. [27] modeled carbon sequestration in Iran’s Nour-Rud watershed, estimating a loss of 9.9 million tons of carbon between 1988 and 2018, with further declines projected by 2048, alongside economic losses exceeding $15 million. Dey et al. [28] simulated urban development scenarios in Kolkata, India, showing a 21% increase in carbon storage under green space restoration and a 10–15% decline under infrastructure-driven growth. Pache et al. [29] quantified carbon sequestration in Romania’s Retezat National Park and validated the economic value of forest-based carbon storage using both management plans and LiDAR data. However, this comprehensive approach remains notably absent in studies of institutional landscapes, particularly in sub-Saharan Africa.

Furthermore, the economic dimension of campus carbon storage also remains primarily unstudied, particularly in developing country contexts where institutional budget constraints and local economic conditions may require different valuation approaches than those typically applied in developed countries. Economic frameworks for carbon valuation in African contexts show promise, with studies estimating sequestration costs and values ranging from $13.30 per ton/ha to under $30 ha⁻¹ yr⁻¹, depending on system type and market conditions [30,31,32].

These combined gaps, (1) the lack of integrated land-use change and carbon storage assessments for institutional landscapes in Sub-Saharan Africa (SSA), (2) the absence of applications of the InVEST model for campus sustainability planning in the region, and (3) the need for context-specific economic valuation of carbon storage in developing economies, limit both scientific understanding and practical management applications. This study addresses these gaps through a comprehensive assessment of land use change, carbon storage dynamics, and economic valuation across three major university campuses in Ghana: University of Ghana (UG), Kwame Nkrumah University of Science and Technology (KNUST), and University of Cape Coast (UCC). The research objectives are to: (1) quantify historical land use changes and carbon storage dynamics from 2017 to 2023; (2) develop and analyze five future land use scenarios using the InVEST carbon model; (3) conduct economic valuation using Ghana-specific social cost of carbon estimates; and (4) provide evidence-based recommendations for sustainable campus planning that balances development needs with carbon storage objectives.

The novelty of this research, therefore, lies in the first integrated application of this comprehensive framework, historical land use change analysis, prospective scenario modeling with the InVEST carbon model, and context-specific economic valuation, to university campus settings in West Africa. This provides both a methodological advancement for campus sustainability science and regionally relevant findings for institutional planning. Our work builds on emerging research demonstrating African universities’ critical role in climate change mitigation and contributes to a growing understanding of carbon dynamics in Ghanaian ecosystems, complementing field-based studies such as [33]. By bridging these strands of research, this study provides a transferable, evidence-based framework for understanding and optimizing the carbon storage potential of institutional landscapes in rapidly urbanizing tropical environments.

The remainder of this paper is structured as follows. Section 2 (Data and Methodology) details the study areas, data sources, and the analytical procedures for land use change detection, carbon assessment with the InVEST model, economic valuation, and scenario development. Section 3 (Results) presents the findings on historical land use change, projected scenarios, carbon storage dynamics, and their economic implications. Section 4 (Discussion) interprets these results in the broader context of regional land change, climate mitigation, and institutional planning, while also considering the study’s limitations. Finally, Section 5 (Conclusions) summarizes the key findings and provides evidence-based recommendations for sustainable campus planning.

2. Data and Methodology

2.1. Study Area

The University of Ghana (UG) is located in Legon, about 12 km northeast of the Accra metropolis, where Ghana’s capital city is situated. The Accra metropolis encompasses a population of 284,124 individuals [34]. The university’s main campus occupies approximately 10 km² and is notable for its lush green environment, characterized by urban forestry elements such as the well-known UG Botanic Gardens, tree-lined pathways throughout the campus, and vegetated areas within residential zones, contributing to its ecological vibrancy. UG’s botanical garden was established 1948 originally to serve the educational and leisure needs of its students. The garden which spans 50 hectares is owned by the University of Ghana and managed by the Department of Plant Biology and Environmental Studies and Mulch Company Limited [35].

The Kwame Nkrumah University of Science and Technology (KNUST) is located the Kumasi metropolitan area of Ashanti Region, Ghana. Within this bustling urban setting of Kumasi, there is a population of 443,981 residents [34]. KNUST supports a vibrant community of students and a workforce contributing to its academic and operational activities. The university has a botanical garden that was established in 1960 and has since been useful in terms of education, research and recreation [36]. KNUST stands out for its rich urban green spaces, featuring an array of forestry resources such as a botanical garden, public parks, trees along avenues, and greenery integrated into residential areas, which collectively enhance its environmental landscape and contribute to academic well-being [37].

The Cape Coast metropolitan area, home to the University of Cape Coast (UCC), covers 122 km² and has a population of 189,925, according to the Ghana Statistical Service [34]. UCC was established in October 1962 as a university college and attained full university status on 1 October 1971. The main campus is situated approximately 5 km west of Cape Coast, the capital of the Central Region. The university has a zoo which hosts a variety of animals [38]. In 2022, UCC was the topmost ranked university in Ghana and West and among the top five universities in Africa [39].

2.2. Spatial Data Acquisition and Processing

Campus boundaries were extracted from OpenStreetMap contributors (2023) using QGIS software version 3.24.1 (https://www.openstreetmap.org/) (Figure 1). Land use data for 2017 and 2023 were obtained from ESRI’s 10 m annual land cover products derived from Sentinel-2 imagery with >75% classification accuracy. The dataset includes nine land cover classes: water, trees, crops, built areas, bare ground, rangeland, flooded vegetation, snow/ice, and clouds [40].

Spatial data processing was conducted in ArcGIS Pro 3.4.0 through sequential steps: (1) campus-specific land use map extraction through shapefile masking; (2) land use area quantification using the Tabulate Area tool; (3) change detection analysis using the Change Detection Wizard to produce transition matrices; and (4) visualization using R language (

Table 1. Data sources and characteristics used in the land use change and carbon storage assessment.

Data Type	Source	Temporal Coverage	Spatial Resolution	Accuracy	Access
Campus Boundaries	OpenStreetMap Contributors (2023)	Current	Vector polygons	User-validated	https://www.openstreetmap.org
Land Cover Maps	ESRI 10 m Annual Land Cover (Karra et al., 2021 [40])	2017, 2023	10 m	>75%	https://livingatlas.arcgis.com/landcover (accessed on 20 January 2025)
Carbon Pool Values	Leh et al. (2013) [41]	-	Ecosystem-specific	Literature-derived	https://doi.org/10.1016/j.agee.2012.12.001
Social Cost of Carbon	Ricke et al. (2018) [42]	Country-specific	National	Model-based	https://country-level-scc.github.io (accessed on 20 January 2025)

).

2.3. Land Use Change Detection and Carbon Storage Assessment

Change detection analysis was performed using ArcGIS Pro’s Change Detection Wizard to systematically compare the 2017 and 2023 land cover classifications. The process involved a pixel-by-pixel comparison to identify land cover transitions and generate comprehensive transition matrices. Tabular outputs were organized using pivot table methodology to quantify transition magnitudes and identify dominant change patterns.

The ESRI 10 m land cover maps served as the primary input, defining the spatial extent and distribution of LULC classes. Carbon storage and sequestration were estimated using InVEST Carbon Storage and Sequestration model (version 3.8.3) [43]. The model operates on a foundational principle: the carbon stored in a landscape is a function of its land use types. The model assumes static land cover and does not simulate dynamic processes such as tree growth or soil carbon fluxes. Scenario projections were developed using the InVEST proximity-based scenario generator, and the resulting land cover maps were used to estimate future carbon storage under each scenario. Each land use class is assigned a total carbon density value, and the model calculates the total carbon stock by multiplying the area of each LULC class by its respective carbon density:

$$Cs=C_a+C_b+C_{soil}+C_{dom}$$

(1)

where $Cs$ is total carbon stored (Mg C), $C_a$ is the aboveground biomass, $C_b$ is the belowground carbon biomass, $C_{soil}$ is soil carbon and $C_{dom}$ is the dead organic matter. The result is in megagrams of carbon per hectare (Mg C).

Carbon pool values were derived from the regional study by Leh et al. [41], which provides regionally appropriate estimates for West African ecosystems. Values from this study were aggregated and mapped to correspond with the ESRI classification scheme. The final carbon density values (in Mg/ha) applied in the model for each land cover class are summarized in

Table 2. Carbon density values (Mg/ha) per land use/land cover class used in the InVEST model.

Land Use Type	Above Ground-Biomass	Below-Ground Biomass	Soil Organic Carbon
Water	0	0	0
Trees	162.9	31.4	14.2
Croplands	62.4	20.6	13.2
Built areas	3.0	0.6	13.5
Bare grounds	2.3	0.5	21.7
Rangelands	23.9	1.0	18.1

below. This approach was necessary due to absence of site-specific measurements and represents an acknowledged limitation.

2.3.1. Economic Valuation Methodology

Economic valuation was performed on the carbon sequestration outcomes estimated by the InVEST model. The calculation employed the model’s net present value (NPV) module (USD in this study), using the carbon storage results from the 2023 land use map as a baseline and the results from the scenario projections as future outcomes:

$${NPV}_{cs}=V{\displaystyle \frac{s_x}{q-p}}\sum _{t=0}^{q-p-1}{\displaystyle \frac{1}{{\left(1+\frac{r}{100}\right)}^t{\left(1+\frac{c}{100}\right)}^t}}$$

(2)

where ${NPV}_{cs}$= Net Present Value of carbon sequestration, V = social cost of carbon ($/tCO₂), $s_x$ = amount of carbon sequestered on land parcel (Mg C), q = future year, p = current year, r = yearly carbon price discount rate and c = yearly change in price of carbon.

SCC is economic damage to society owing to an increase in a ton of carbon or the profit from reducing a ton of emitted carbon. The social cost of carbon value was derived from Ghana-specific estimates by [42] of $0.970/tCO₂ (https://country-level-scc.github.io/explorer/ accessed on 30 January 2025). While greenhouse gases are global pollutants with impacts independent of emission location, the use of a country-specific SCC is justified for several reasons relevant to local decision-making. First, it reflects the local economic conditions and national climate vulnerability, representing the marginal economic damage to Ghanaian society from a ton of emissions or the benefit of its abatement [42,44]. Second, for national and sub-national institutions like universities, a global SCC value (often ranging from $50–200/tCO₂) would yield economic valuations that are orders of magnitude larger than local budget constraints and opportunity costs, making the results less actionable for practical campus planning [45]. Using a localized estimate provides a more realistic and context-appropriate basis for cost–benefit analysis and trade-off evaluation at the institutional level in Ghana, where resource allocation decisions are made based on local economic realities. While substantially lower than the $50–200/tCO₂ range commonly used in international climate policy assessments, these localized estimates offer more appropriate guidance for institutional decision-making where global carbon prices may not reflect budget constraints and opportunity costs [40].

This valuation approach draws on foundational economic theory, beginning with [46], who argued that discounting future utility is ethically indefensible and stems from a lack of imagination. Portney (1994) [47] emphasized that moral reasoning requires treating future generations equitably, advocating for intergenerational fairness in policy design. A 0% discount rate is therefore included to reflect a pure intergenerational equity perspective, assuming that the welfare of future generations should be valued equally to that of the present [48]. In climate policy and cost–benefit analysis, a 3% discount rate is commonly used to represent the social rate of time preference, reflecting the after-tax return on household savings and balancing present and future consumption values [49,50]. Conversely, a 7% discount rate is applied to reflect the opportunity cost of capital, representing the average pre-tax return on private investments and typically used in infrastructure-focused evaluations where capital displacement is a concern [28,51]. By incorporating all three rates, this study captures a range of ethical and economic perspectives on carbon sequestration valuation, offering a comprehensive framework to inform sustainable campus planning in Ghana.

2.3.2. Scenario Development and Analysis

Five land use scenarios were developed using the InVEST proximity-based scenario generator to evaluate alternative campus development pathways and their impact on carbon storage. The tool requires the specification of (1) focal land use (areas where change originates), (2) convertible land use (areas available for conversion), and (3) replacement land use (the new land use type). Conversion patterns are determined by proximity algorithms, either “nearest to edge” to simulate contiguous expansion (e.g., urban sprawl) or “furthest from edge” to simulate fragmented, interior conversion (e.g., agroforestry integration).

Scenario parameters were selected based on observed historical change patterns (2017–2023) and practical considerations for campus planning in the Ghanaian context. The scenarios were designed to represent realistic policy choices faced by university administrators. The five scenarios were developed to reflect plausible campus development pathways based on observed land use trends and practical planning considerations. Each scenario simulates land cover changes over a 10-year horizon. Infrastructure Development (InfraDev) represents continued infrastructure expansion, while Agroforestry (AgroFor) and Tree Expansion (TreeExp) simulate moderate ecological restoration through agroforestry and tree planting. Extensive Tree Expansion (ExtTreeExp) reflects an ambitious restoration effort, and Agroforesty (AgroPlus) combines agroforestry and tree expansion strategies. These scenarios were designed to explore trade-offs between development and conservation, providing insights into the carbon storage potential of different planning approaches. The five scenarios developed are detailed below:

• InfraDev: This scenario simulates continued campus expansion through the conversion of natural areas to built infrastructure, a common pressure on Ghanaian universities. Forest and rangeland areas were designated as convertible land uses, with built areas as the replacement. Existing built areas were set as the focal land use. Maximum conversion was set at 10% of the existing built area coverage to reflect typical institutional expansion rates. Conversion followed the “nearest to edge” algorithm, simulating development sprawling outward from existing infrastructure.
• AgroFor: This scenario models the integration of trees into existing agricultural areas to enhance sustainability. Cropland was designated as both the focal and convertible land use, with forest as the replacement. A maximum of 10% of total cropland was set for conversion. Conversion followed the “furthest from edge” algorithm with 10 conversion steps to create intentionally fragmented patterns typical of real-world agroforestry systems where trees are dispersed within crop fields.
• TreeExp: This scenario represents a moderate, targeted effort to increase tree cover. A total of 5% of the combined area of cropland and rangeland was set for conversion. These land uses were designated as convertible, with forest as the replacement. Conversion followed the “nearest to edge” algorithm to simulate natural forest expansion patterns and the consolidation of green spaces.
• ExtTreeExp: This scenario implements an ambitious tree cover expansion policy, doubling the effort of the TreeExp scenario. A total of 10% of the combined crop and rangeland areas were set for conversion to forest. The configuration (convertible land use, replacement, algorithm) otherwise follows the TreeExp methodology to represent a more intensive restoration effort.
• AgroPlus: This combined scenario integrates both agroforestry and general tree expansion strategies to represent a comprehensive green infrastructure policy. It simultaneously implements the conversions defined in both the AgroFor (10% of cropland to forest) and TreeExp (5% of combined crop/rangeland to forest) scenarios. This approach tests the potential synergistic benefits of applying multiple restoration strategies across the campus landscape.

Previous studies have successfully applied the InVEST proximity-based scenario generator to model land use and ecosystem service changes in diverse contexts, including carbon storage in [28], heat mitigation in Nagpur [52] and ecosystem service trade-offs in Odisha’s mangrove region [53], validating its utility for this assessment.

3. Results

3.1. Historical Land Use Changes and Carbon Storage Dynamics (2017–2023)

All three campuses experienced significant land use changes dominated by built area expansion and tree cover decline between 2017 and 2023 (Figure 2).

From Figure 3, built areas increased from 571.03 to 676.94 hectares (+105.9 ha, +18.5%) at UG, from 533.1 to 610.9 hectares (+77.8 ha, +14.6%) at KNUST, and from 289.52 to 302.88 hectares (+13.4 ha, +4.6%) at UCC. Built area coverage increased from 45.6% to 53.9% (UG), 53.3% to 61.1% (KNUST), and 58.6% to 61.3% (UCC).

Tree cover declined across all campuses, with KNUST experiencing the most severe loss of 120.9 hectares (−52.3%), from 231.3 to 110.4 hectares. UG lost 103.5 hectares (−23.2%), declining from 445.2 to 341.7 hectares, while UCC showed the smallest loss of 6.4 hectares (−4.1%), from 155.2 to 148.8 hectares. Rangeland dynamics varied across campuses. KNUST and UCC experienced expansion (10.0 ha and 28.6 ha, respectively), while UG showed modest decline (−4.7 ha). Cropland increased at UG (3.0 ha) and expanded substantially at KNUST (+35.9 ha), while UCC experienced reduction (−35.5 ha). Water cover remained relatively stable across all campuses.

Land use transition analysis revealed high persistence of built areas (>96% unchanged across all campuses) and significant conversion of tree cover to other land uses (Figure 4). At KNUST, 55.8% (129.1 ha) of original tree cover was converted, primarily to built areas (57.2 ha) and rangelands (57.5 ha). UG converted 36.5% (160.8 ha) of tree cover, mainly to built areas (75.4 ha) and rangelands (70.1 ha). UCC demonstrated the greatest tree cover retention, converting only 19.3% (29.9 ha), primarily to built areas (12.4 ha) and rangelands (17.3 ha).

These transitions reflect the intensity of development pressures and the varying effectiveness of campus-level land management strategies. The high conversion rates of tree cover to built areas and rangelands underscore the need for strategic planning to preserve ecological functions. The quantified land use transitions across UG, KNUST, and UCC demonstrate the extent of environmental change over the six-year period, with KNUST showing the highest proportional conversion of tree cover and UCC exhibiting the strongest retention.

The quantified land use transitions and carbon losses across UG, KNUST, and UCC demonstrate the extent of environmental change over the six-year period, with KNUST showing the highest proportional reduction in carbon storage.

3.2. Land Use Changes Under Scenario Projections

The five scenarios produced distinct land use changes from the 2023 baseline, with the magnitude and direction of change for key classes summarized in

Table 3. Percentage change in land use area from the 2023 baseline under the five projected scenarios.

Campus	Land Use Type	InfraDev	AgroFor	TreeExp	ExtTreeExp	AgroPlus
UG	Trees	−9.1	1.4	5	10	6.4
	Crops	0	−10.1	−8.5	−17.5	−25.4
	Built areas	10	0	0	0	0
	Rangelands	−19.7	0	−7.1	−14	−5.3
KNUST	Trees	−15.2	4.2	5	10	9.2
	Crops	0	−9.9	−3.2	−5.5	−15.2
	Built areas	10	0	0	0	0
	Rangelands	−19	0	−17	−3.6	−1.3
UCC	Trees	−13.9	0.3	5	10	5.3
	Crops	0	−9.7	35	−58.4	−65
	Built areas	9.9	0	0	0	0
	Rangelands	−24.9	0	−15.1	−32	−12.3

. As designed, the Infrastructure Development (InfraDev) scenario resulted in a uniform ~10% expansion of built-up areas across all campuses, directly driving proportional losses in natural land covers. Tree cover declined most severely at KNUST (−15.2%), followed by UCC (−13.9%) and UG (−9.1%). Rangelands were also significantly reduced, with the highest loss observed at UCC (−24.9%).

In contrast, greening scenarios had minimal impact on built areas but substantially altered vegetative covers. The AgroFor scenario, targeting cropland, resulted in modest tree gains (+0.3% to +4.2%) and corresponding cropland losses (~−10%). The targeted tree expansion scenarios (TreeExp and ExtTreeExp) successfully increased tree cover by the intended 5% and 10%, respectively. This was achieved primarily through the conversion of rangeland and, most notably, cropland at UCC, which saw cropland reductions of −35.0% and −58.4%. The combined AgroPlus scenario generated the most complex changes, yielding tree cover increases between 5.3% and 9.2% while dramatically reducing cropland area, particularly at UCC (−65.0%). These results clearly illustrate the land-use trade-offs inherent in each campus development strategy.

3.3. Changes in Carbon Storage, Actual Land Use and Scenario-Based Projections

Historical land use changes between 2017 and 2023 resulted in substantial carbon storage losses across all campuses, with UG experiencing the largest absolute loss (−19,695 Mg C) and KNUST the largest proportional loss (−29.5%) from its 2017 stock (

Table 4. Historical and projected carbon storage (Mg C) under different land use scenarios. The baseline years are 2017 and 2023. Five future scenarios were modeled using the InVEST Carbon model: Infrastructure InfraDev, AgroFor, TreeExp, ExtTreeExp, AgroPlus. The bottom section shows the change from the 2023 baseline.

Campus	2017	2023	InfraDev	AgroFor	TreeExp	ExtTreeExp	AgroPlus
UG	115,019	95,325	88,434	95,853	97,945	100,551	98,303
KNUST	67,990	47,927	43,561	48,444	48,759	49,613	49,224
UCC	41,589	38,297	34,086	38,354	39,426	40,604	39,427
Change from 2023
UG			−6891	+528	+2620	+5227	+2978
KNUST			−4366	+517	+832	+1686	+1296
UCC			−4211	+56	+1129	+2307	+1130

).

Projections under the five scenarios revealed significant variations in future carbon outcomes (Table 4). The InfraDev scenario projected substantial additional losses, further reducing carbon storage by 4211–6891 Mg C across the campuses. In contrast, greening scenarios led to gains. The AgroFor scenario showed the most modest benefits (+56 to +528 Mg C). More ambitious interventions yielded stronger results: the TreeExp and AgroPlus scenarios generated intermediate gains (+832 to +2978 Mg C), while the ExtTreeExp scenario offered the highest potential for carbon recovery, adding between 1686 Mg C (KNUST) and 5227 Mg C (UG).

Crucially, even the most ambitious scenario (ExtTreeExp) could not fully restore carbon stocks to 2017 levels within the projection period. UG would recover to 87.4% of its 2017 stock, KNUST to 73.0%, and UCC, demonstrating the greatest resilience, to 97.6%. This underscores the long-term consequence of historical land-use decisions and the time needed for ecosystem recovery.

3.4. Economic Valuation of Carbon Sequestration Services

From Figure 5, the economic analysis revealed substantial variation in NPV across scenarios and discount rates. These negative values indicate reduction in NPV. The ExtTreeExp scenario generated the highest NPVs: $5070/$5040/$5000 at UG, $1640/$1630/$1610 at KNUST, and $2240/$2220/$2210 at UCC across 0%/3%/7% discount rates, respectively.

Conversely, the InfraDev scenario resulted in substantial economic losses: −$6680/−$6640/−$6590 at UG, −$4240/−$4210/−$4180 at KNUST, and −$4090/−$4060/−$4030 at UCC across the three discount rates. These negative values indicate significant economic costs of continued infrastructure expansion through carbon storage losses. Moderate tree expansion scenarios generated positive but lower NPVs. TreeExp produced $2540/$2530/$2510 (UG), $810/$800/$800 (KNUST), and $1100/$1090/$1080 (UCC). AgroPlus yielded $2890/$2870/$2850 (UG), $1260/$1250/$1240 (KNUST), and $1100/$1090/$1080 (UCC). AgroFor showed the smallest economic benefits with $510/$510/$510 (UG), $500/$500/$490 (KNUST), and $55/$55/$55 (UCC). NPV decreased consistently with higher discount rates across all scenarios, though the differences were modest (1–2%), with differences of 1–2% between discount rates. The economic differential between ExtTreeExp benefits and InfraDev costs ranged from $5850 to $11,750 across campuses and discount rates.

Net present value calculations indicate that tree-based scenarios yield positive economic returns under all discount rates, while infrastructure expansion scenarios result in consistent economic losses.

3.5. Cross-Campus Performance and Institutional Factors

Comparative analysis revealed distinct performance patterns across the three campuses that reflect varying institutional contexts and development pressures (

Table 5. Comparative summary of land area, carbon loss, restoration potential, and economic valuation across three Ghanaian university campuses (UG, KNUST, and UCC) from 2017 to 2023. The table presents key metrics including total and annual carbon loss, percentage loss rates, projected carbon gains under the best-performing scenario, and NPV ranges based on a 0% discount rate.

Metric	UG	KNUST	UCC
Campus Area (total ha)	1255.6	1000.5	493.9
Historical Carbon Loss (2017–2023)
Total loss (Mg C)	−19,695	−20,063	−3292
Annual loss rate (Mg C/year)	−3282	−3344	−549
Loss rate (% per year)	−2.9%	−4.9%	−1.3%
Restoration Potential
Best scenario gain (Mg C)	+5227	+1686	+2307
% improvement over 2023	+5.5%	+3.5%	+6.0%
Economic Performance (0% discount rate)
Best scenario NPV (USD)	+5070	+1630	+2240
Worst scenario NPV (USD)	−6690	−4240	−4080
Economic range (USD)	11,750	5870	6320

). UG demonstrated the largest absolute environmental impacts in terms of both historical losses (19,695 Mg C) and potential gains (5227 Mg C under ExtTreeExp), reflecting its large total area (1254 ha) and substantial baseline tree cover.

Annual carbon loss rates during the study period averaged 3282 Mg C/year at UG.

KNUST exhibited the most severe proportional environmental degradation with 29.5% historical carbon losses and continued vulnerability under infrastructure scenarios. The campus also showed the most limited restoration potential, achieving only 73.0% of 2017 carbon levels under the most ambitious tree expansion scenario. This pattern indicates particularly intense development pressures and constraints on environmental recovery.

UCC demonstrated the greatest environmental resilience with only 7.9% historical carbon losses and strong restoration potential, reaching 97.6% of 2017 carbon levels under ExtTreeExp. The campus also showed the highest carbon storage density per unit area and the most effective tree cover preservation during the study period.

Economic performance patterns followed environmental outcomes, with UG generating the largest absolute economic benefits ($5070 under ExtTreeExp at 0%) but also facing the highest economic costs (−$6690 under InfraDev). KNUST showed modest economic potential ($1635 under ExtTreeExp) constrained by historical environmental degradation. UCC achieved intermediate economic outcomes ($2240 under ExtTreeExp) with strong potential relative to campus size. Built area persistence exceeded 96% across all campuses, confirming that infrastructure development represents largely irreversible land use change. Tree cover conversion rates varied substantially, with 55.8% conversion at KNUST, 36.5% at UG, and 19.3% at UCC, indicating different vulnerabilities to development pressure.

The relationship between campus size, urban context, and environmental outcomes varied across institutions. UCC’s superior environmental performance despite substantial built coverage (61.3%) suggests that institutional policies and management approaches significantly influence environmental outcomes beyond simple area or urban pressure effects. KNUST’s severe environmental degradation despite botanical garden presence indicates that institutional environmental commitments require comprehensive implementation across entire campus areas to be effective.

3344 Mg C/year at KNUST, and 549 Mg C/year at UCC, equivalent to 12,035, 12,261, and 2013 tCO₂ annually, respectively. These rates exceed typical deforestation rates in many natural forest systems and highlight the intensity of development pressures on institutional landscapes. The scenario analysis indicates that policy interventions could substantially alter environmental trajectories, with the difference between worst-case (InfraDev) and best-case (ExtTreeExp) scenarios representing carbon storage variations of 12,118 Mg C (UG), 6052 Mg C (KNUST), and 6518 Mg C (UCC). Comparative metrics across campuses highlight variation in carbon loss rates, restoration potential, and economic outcomes, reflecting differences in campus size, baseline vegetation, and land use dynamics.

4. Discussion

4.1. Drivers and Patterns of Land Use Change on Campuses

The results of this study reveal that university campuses in Ghana are undergoing rapid and substantial land use transformations. Built area expansion rates of 14.6–18.5% over six years significantly exceed the typical urban growth rates of 3–5% annually reported for sub-Saharan Africa [54]. This suggests that institutional landscapes are subject to development pressures that are more intense than those affecting general urban environments. These findings are consistent with previous studies documenting the sprawling nature of urban expansion in Accra and Kumasi, where forest and agricultural lands are increasingly converted to built infrastructure [55,56].

Tree cover losses of 23–52% across the campuses are particularly striking. These rates surpass those documented in similar tropical university studies from Southeast Asia (10–20%) and Latin America (5–15%), indicating that West African institutional landscapes may be more vulnerable to environmental degradation [57,58]. The contrast between UCC’s minimal tree cover loss (4.1%) and KNUST’s severe degradation (52.3%) underscores the role of institutional management practices in shaping environmental outcomes. This finding supports the view that effective stewardship can mitigate the adverse effects of urbanization, even within development-oriented institutions.

Built area persistence exceeding 96% across all campuses confirms that infrastructure development represents a largely irreversible form of land use change. This observation aligns with the broader literature on urban land transformation, which emphasizes the permanence of built environments and the difficulty of restoring ecological functions once natural cover is lost. Therefore, it is important to recognize that proactive planning and strategic conservation are essential for maintaining ecosystem services in institutional settings.

4.2. Carbon Losses and Climate Mitigation Potential

The carbon storage losses documented in this study, totaling 43,050 Mg C across UG, KNUST, and UCC, represent a significant environmental impact. These losses are equivalent to approximately 158,000 tCO2, which corresponds to the annual emissions of over 34,000 passenger vehicles or the deforestation of 350 hectares of tropical forest. Such figures highlight the importance of including institutional land management in national climate accounting and mitigation strategies.

Carbon storage densities observed in this study (248 Mg C/ha) substantially exceed those reported from North American university campuses (50–150 Mg C/ha), reflecting the higher sequestration potential of tropical vegetation and soils [59,60]. Studies from tropical urban forests report similar densities (218–236 Mg C/ha) but much lower loss rates (5–10%) over comparable periods, suggesting that West African institutional landscapes face more severe development pressures than similar institutions in other tropical regions [61,62].

Ref. [23] reported average carbon stock densities of 46.77 tons/ha across 17 university campuses in Bangkok, which are notably lower than the values observed in this study. This comparison underscores the higher carbon sequestration potential of West African vegetation and supports the argument that institutional landscapes in Ghana represent valuable carbon sinks. Additional studies in Kumasi’s urban green spaces report carbon stocks of 270 ± 22 t C/ha in designated green areas [18], 228 t C/ha across the city’s urban forest [63], and 53.9 ± 11.7 Mg C/ha in public parks [64], further validating the ecological significance of campus environments.

Scenario modeling revealed that even under the most ambitious tree expansion scenario (ExtTreeExp), none of the campuses would fully recover their 2017 carbon storage levels within the 10-year projection period. UG would reach 87.4% of 2017 levels, KNUST 73.0%, and UCC 97.6%. These findings are consistent with restoration ecology literature, which suggests that tropical forest recovery requires 20–50 years to approach mature carbon storage levels [65]. This is consistent with findings from Kumasi, where urban land transformation has significantly reduced tree cover [66], therefore prompting the need to explore nature-based solutions to restore ecosystem services.

4.3. The Economic Case for Carbon-Conscious Planning

The economic valuation conducted in this study provides compelling evidence for the financial viability of conservation-oriented campus planning. Tree expansion scenarios generated positive net present values (NPV) ranging from $1610 to $5070, while infrastructure-focused development resulted in economic losses between $4030 and $6690. These results support the hypothesis that environmental investments can yield tangible economic benefits, even under conservative valuation assumptions.

The use of Ghana-specific social cost of carbon ($0.970/tCO2) reflects local economic conditions and climate vulnerability [38]. While this value is substantially lower than the $50–200/tCO2 range commonly used in international climate policy assessments, it provides a realistic basis for institutional decision-making. When moderate international SCC estimates are applied, the economic benefits of tree expansion increase dramatically, up to $261,000 at UG, $84,000 at KNUST, and $115,000 at UCC, suggesting strong potential for external financing through carbon credit mechanisms or international climate funds.

These findings have important implications for institutional budgeting and sustainability planning. Universities could prioritize investments in tree planting, agroforestry integration, and green infrastructure as part of their climate action strategies. Such initiatives may also enhance eligibility for climate-related funding and contribute to national and regional mitigation targets. A reasonable approach to tackle this issue could be to establish carbon accounting frameworks and integrate ecosystem service valuation into campus master plans.

4.4. Lessons from Cross-Campus Comparisons for Institutional Management

The comparative analysis across UG, KNUST, and UCC reveals distinct performance patterns that reflect varying institutional contexts and development pressures. UG demonstrated the largest absolute environmental impacts, with historical carbon losses of 19,695 Mg C and potential gains of 5227 Mg C under ExtTreeExp. KNUST exhibited the most severe proportional degradation, losing 29.5% of its carbon storage and achieving only 73.0% of 2017 levels under the best-case scenario. This pattern suggests that development pressures at KNUST are particularly intense and that restoration potential is constrained by limited remaining natural cover.

UCC, by contrast, demonstrated the greatest environmental resilience, with only 7.9% historical carbon loss and restoration potential reaching 97.6% of 2017 levels. The campus also showed the highest carbon storage density per unit area and the most effective tree cover preservation during the study period. These findings suggest that institutional policies and management approaches significantly influence environmental outcomes beyond simple urban pressure effects.

Annual carbon loss rates, 3282 Mg C/year at UG, 3344 Mg C/year at KNUST, and 549 Mg C/year at UCC, exceed typical deforestation rates in many natural forest systems, highlighting the intensity of development pressures on institutional landscapes. The difference between worst-case (InfraDev) and best-case (ExtTreeExp) scenarios represents carbon storage variations of 12,118 Mg C (UG), 6052 Mg C (KNUST), and 6518 Mg C (UCC), demonstrating the critical importance of strategic campus planning for environmental outcomes.

4.5. A Framework for Sustainable Campus Planning and Policy

The methodological framework developed in this study, combining historical analysis, scenario modeling, and economic valuation, offers a transferable tool for environmental assessment and planning in institutional contexts. University campuses can serve as demonstration sites for urban sustainability, providing technical expertise, policy innovation, and leadership in climate action [57].

To integrate environmental considerations into institutional planning, policy frameworks could draw inspiration from successful models in other developing countries. These may include carbon impact assessment requirements, incentives for green infrastructure investment, and the adoption of environmental performance standards. Regional cooperation among institutions could further enhance effectiveness through knowledge sharing, technical collaboration, and economies of scale in implementation.

Overall, the findings suggest that institutional landscapes are an underutilized opportunity for climate mitigation and ecosystem service enhancement. By adopting evidence-based planning approaches, universities can contribute meaningfully to national sustainability goals while improving campus livability and resilience.

4.6. Study Limitations and Future Research Directions

While this study provides a comprehensive assessment of carbon storage dynamics on university campuses in Ghana, several limitations should be acknowledged. The land use maps used have a classification accuracy of 75%, which may introduce errors in carbon storage calculations, particularly for transitions between rangelands and croplands. Carbon pool values were derived from [41], which, although regionally appropriate, may not fully reflect site-specific conditions.

The InVEST model does not account for dynamic processes such as soil carbon emissions from land use changes or long-term growth of newly planted trees. Scenario parameters were selected intuitively, based on historical patterns and practical considerations, rather than through stakeholder consultation. This limits the applicability of the scenarios to real-world planning contexts.

Future research should address these limitations by incorporating field-based carbon measurements, dynamic modeling of carbon fluxes, and participatory scenario development. Expanding the range of economic parameters and conducting sensitivity analyses would also enhance the robustness of valuation outcomes. Further investigation into the role of institutional governance, stakeholder engagement, and policy implementation could provide valuable insights for scaling up campus sustainability initiatives.

5. Conclusions

This study provides the first comprehensive assessment of land use change, carbon storage dynamics, and economic valuation across university campuses in Ghana. The findings reveal that institutional landscapes are undergoing rapid transformation, with built area expansion and tree cover loss significantly exceeding regional urbanization trends. Between 2017 and 2023, UG, KNUST, and UCC collectively lost 43,050 Mg C, with KNUST experiencing the most severe proportional decline (29.5%) and UCC demonstrating the greatest resilience (7.9%).

Scenario modeling demonstrated that infrastructure-focused development would result in additional carbon losses of 4211–6891 Mg C, while tree expansion scenarios could increase storage by 1686–5227 Mg C. However, even under the most ambitious restoration scenario, full recovery of 2017 carbon levels will not be achieved within the 10-year projection period. This is consistent with findings from Kumasi, where urban land transformation has significantly reduced tree cover [62], prompting the need to explore nature-based solutions to restore ecosystem services.

Economic valuation using Ghana’s social cost of carbon ($0.970/tCO2) revealed that tree expansion scenarios generate positive net present values ($1610–$5070), while infrastructure expansion imposes substantial costs ($4030–$6690). These results underscore the economic viability of conservation-oriented planning and suggest opportunities for external financing through carbon credit mechanisms and climate funds.

Cross-campus comparisons highlighted the influence of institutional management on environmental outcomes. UG exhibited the largest absolute impacts, KNUST faced the greatest proportional degradation, and UCC demonstrated strong resilience and restoration potential. These differences emphasize the need for campus-wide integration of sustainability practices and strategic planning.

The methodological framework developed in this study, combining historical analysis, scenario modeling, and economic valuation, is transferable to other institutional contexts. University campuses can serve as demonstration sites for urban sustainability, contributing to national climate goals and regional cooperation.

We recommend that universities incorporate systematic carbon storage assessments into planning processes, establish measurable environmental targets, and pursue collaborative strategies for restoration and conservation. Future research should include field-based carbon measurements, dynamic modeling of carbon fluxes, and stakeholder-informed scenario development to enhance the robustness and applicability of campus sustainability initiatives.