https://orcid.org/0000-0002-7593-1446
Figures
Abstract
Ecosystem change affects the availability of resources and services provided by nature. Ecosystem Natural capital accounting helps track these changes and supports better decision-making for managing the environment. This approach aims to assess changes in the stocks and flows of natural resources and the possibility to integrate them into economic and political decisions. The protected area of Mahavavy-Kinkony Complex, in North-Western of Madagascar, was chosen to implement this approach due to its many types of ecosystems as well as important reserves of threatened birds. In five years (2013–2018), we have observed a reduction in woodland cover (forest and mangrove) due to both regulated and illegal logging, linked to urban expansion and increasing of human pressure. This loss of woodland compromises not only biodiversity but also the capacity of ecosystems to provide ecosystem services. At the same time, the silting up of surface waters is compromising water quality and the health of aquatic ecosystems. In addition, the increase in agricultural land at the expense of forested areas raises concerns about the continuing degradation of natural ecosystems. All of these changes can be observed inside local socio-ecological landscape type. Each socio-ecological landscape type shows the potential variation in the production of ecosystem services.
Citation: Ramihangihajason TA, Weber J-L, Rakotondraompiana S, Roger E, Faramalala MH, Rakotoniaina S (2025) Natural capital accounting as a decision support tool for environmental management of a protected area in Madagascar. PLoS One 20(5): e0321948. https://doi.org/10.1371/journal.pone.0321948
Editor: Tzen-Yuh Chiang,, National Cheng Kung University, TAIWAN
Received: September 9, 2024; Accepted: March 13, 2025; Published: May 9, 2025
Copyright: © 2025 Ramihangihajason et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data has been deposited on Zenodo at the following DOI: https://doi.org/10.5281/zenodo.14272482. The Methods section is also available on Zenodo at the following DOI: https://zenodo.org/records/14272568.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
1. Introduction
Awareness of the importance of ecosystems and concerns about their degradation due to human activities has become increasingly recurrent since the middle of 20th century. The first alarm was triggered by the creation of the Red List of Threatened Species by IUCN [1,2]. In 1972, the Club of Rome commissioned the Meadows Report on the Limits to Growth [3]. Then, the Stockholm conference in 1972 [4] highlighted this growing awareness and resulted in the creation of the United Nations Environment Programme (UNEP). The main idea put forward at this conference was to encourage ecologically sound management of the environment. But it was in the Brundtland report [5] that the notion of sustainable development emerged. The sustainable development paradigm underpinned the UN Conference on Environment and Development (UNCED), also known as the ‘Earth Summit’, held in Rio de Janeiro, in 1992 [6]. Outcomes of the Conference are the creation of the 3 conventions on climate change (UNFCCC), desertification (UNCCD) and biological diversity (UNCBD) and to the endorsement of the Agenda 21 [7]. Among other recommendations, the Agenda 21 called for the implementation of natural resource accounts in order to supplement the conventional National Accounts. It gave rise to the publication by UN Statistical Commission (1993) of the first manual of economic-environmental accounting, the so-called SEEA 93. The many experiments of the SEEA lead to the enlargement of its scope. In particular, the role of ecosystems and their services in sustainable development proposed first by Ehrlich and Mooney (1983) [8] was taken into account. The revision of the SEEA decided by the UN Statistical Commission in 2012 resulted in two volumes, the first one called System of Environmental Economic Accounting - Central Framework (SEEA-CF), the second one Ecosystem Accounting (SEEA-EA) [9]. While the SEEA-CF is framed by the classifications of industries and commodities of the System of National Accounts, the SEEA-EA is based on spatial mapping of ecosystems. During the SEEA revision process, the Secretariat of the CBD to fulfil the second objective of the strategic goal A in the Aichi Targets [10], published a technical report called: “Ecosystem Natural Capital Accounts: A Quick Start Package” (ENCA-QSP). It’s aims to implement Aichi Biodiversity Target 2 on Integration of Biodiversity Values in National Accounting Systems in the context of the SEEA Experimental Ecosystem Accounts [11]. While the SEEA-EA includes accounts in physical terms and monetary accounts, ENCA QSP focuses in only biophysical accounts. This initial version concerns only terrestrial ecosystems. In 2021 the SEEA-EA was adopted as an international statistical standard by the UN Statistical Commission [12] subject to monetary accounts being further validated. In the realm of biophysical accounts, SEEA EA and ENCA-QSP are broadly compatible.
Since then, a number of works have been carried out along two lines: implementation of accounts in biophysical terms and monetary valuations. Partial or comprehensive ecosystem accounts physical stocks and flows have been produced for a number of countries using international databases. Accounts targeting to monetary valuation of ecosystem services and assets are developed though case studies in programmes such as the Wealth Accounting and the Valuation of Ecosystem Services (WAVES program) [13] now known as the GPS [14], ARIES for SEEA (Artificial Intelligence for Environment & Sustainability) [15]; CARE-TDL or the Comprehensive Accounting in Respect of Ecology, Triple depreciation line [16]; INVEST or Integrated valuation of ecosystem services and trade-offs [17].
Environmental accounts have primarily been envisaged as an extension of the System of National Accounts, in order to go beyond the GDP (Gross Domestic Product) aggregate and adjust it from hidden environmental costs (negative externalities) or/and unrecorded benefits (positive externalities). In the evolution process of environmental accounting, it first appeared that accounts in biophysical terms were a prerequisite, needed prior to any monetary valuation. As well, the scale of accounting became of growing importance, both considering the analysis of environmental assets and the needs of users, would they be in charge of land management at various levels or companies willing to face their ecological liability.
The feasibility of nation-wide ecosystem accounting has been demonstrated [18–20], their interest as well as their limitations resulting from the use of international datasets. The challenge is now to develop accounts at smaller scales corresponding to more operational applications. In this study, we aim to show that ENCA can be applied at the local scale using high-resolution data and how they can be used as an effective environmental management tool.
In the first section of this paper, the ENCA approach will be summarised. Then, the study area will be described. After that, the results, land cover maps and the five ENCA accounts are presented. The interpretations of these accounts are given in order to obtain a large information useful for environmental managers. Discussions and conclusion will be the last part of this paper.
2. The ecosystem natural capital accounting (ENCA)
The following is mainly taking from Weber (2014) [11] which explain the ENCA terrestrial approach. The detailed methodology is available at this link: https://zenodo.org/records/14272568. The priority of ecosystem natural capital accounting (ENCA) is to show the ecological value of a given place and how ecosystems are evolving. The ecological value is a non-monetary assessment of integrity, health or resilience of ecosystems. All of which are important indicators for determining critical thresholds and minimum requirements to provide ecosystem services [21]. Over time, this ecological value of the study area can decrease, meaning the creation of ecological debts, or can increase meaning an accumulation of ecological credits.
ENCA ecosystem accounts are based on the land cover account which provides the spatial pattern common to other accounts. This is a diachronic analysis of land cover based on two Land Cover maps at the opening date and the closing date. Three ecosystems’ accounts are rooted on this Land Cover account: the ecosystem carbon account, the ecosystem water account and the ecosystem infrastructure integrity and services account. These accounts have the same structure: a set of four tables structured in the same way. The first one shows the basic balance sheet, which includes all ecosystem resource stocks and flows, natural and/or due to human activities between opening and closing dates. The second table measures the ecosystem potential resource and the amount accessible to human use due to constraints (e.g., nature protection). The third one contains all the uses of resources during the time intervals between opening and closing dates. The last table aims at calculating an sustainable use index (the ratio between accessible resources and used resources, compiled in previous parts) and an index of resilience or health of natural resources (a diagnostic based on variables considered as symptoms). The multiplication of sustainable use and health indexes is called ecosystem internal unit value. The average of the three indices of ecosystem internal unit value is then multiplied by the amount of accessible water resources, by the amount of accessible carbon resources and by the amount of accessible infrastructure functional services resources. These last values are the ecosystem capability unit (ECU) for each ecosystem type. Ecosystem capabilities has the status of ecological currency. Then the sum of all ecological values is the Total Ecosystem Capability (TEC) value. All calculation are performed inside analysis units called socio-ecological landscape units (SELUs) [22].
2.1 The socio-ecological landscape unit
Socio-ecological landscape units are areas within which all data are compiled and ecosystem accounts are calculated. SELUs integrate terrestrial and aquatic dimensions. Each SELUs is characterized by the dominant land cover and the boundaries of the sub-watershed.
2.2 The land cover account
The first step of producing the land cover account is to define all land cover classes. Weber (2014) [11] proposed the land cover system units (see Table 1) inspired from Di Gregorio et al. [23]. Sub-classes can be derived from as needed. This nomenclature is compatible with other nomenclature such as the Corine Land cover or others [24–26].
The dates of each image used in this study are for year 2013: 29 April for wet season and 03 August for dry season; for 2018 25 March for wet season, 04 October for dry season.
Each year’s images are actually made up of one Landsat image taken during the dry season and another Landsat image taken during the rainy season. Each Landsat 8 images include 11 spectral bands [27]. However, we only used the 6 following bands: the B2 (blue), B3 (green), B4 (red), B5 (near infrared), B6 (Short-Wave Infrared) and B7 (Short-Wave Infrared). All of these bands have a spatial resolution of 30m x 30m. To each of these spectral bands, we applied an image fusion algorithm [28] with B8 band (panchromatic) which has a spatial resolution of 15mx15m. The objective of image fusion step is to obtain synthetic images with 15mx15m spatial resolution but keeping the original spectral channels.
For each year, the two images (dry season image + rainy season image) are then stacked to form a single image with 15mx15m spatial resolution and 12 spectral bands.
Image of each year is classified using the supervised random forest method [29]. A supervised image classification method requires training and control pixels in a proportion of 70% for training and 30% to control the results. The confusion matrix gives the accuracy of classification.
The choice of training parcels depends on observations made during field work and the spectral responses of image. After classification made, hierarchical randomly points are chosen [30].
The Land Cover changes map can be derived directly from these two Land Cover maps [31], or can be obtained from the spatial image processing using a change detection algorithm to both images [32,33]. All the changes observed will be categorised into change flow classes according to their causes (see Table 2).
2.3 The ecosystem water account
The water ecosystem account includes all continental waters and those of terrestrial ecosystems: water in rivers, canals, lakes and reservoirs, water in soil and vegetation, snow and glaciers. It also contains groundwater in its relationship with surface water. Exchanges with the atmosphere, such as precipitation and real evapotranspiration, are also taken into account. The same goes for water imported into the study area in form of beverage or other forms. User water system is connected to the resource system (canals and pipes, reservoirs) and wastewater returns are considered as secondary resources. In contrast, the account does not cover water stocks in the oceans and subsoil. It only records their relationship with terrestrial ecosystems and continental waters.
Databases on lakes and reservoirs can be used alone or combined with local data to obtain values of depth. For rivers, flows and lengths data also come from international database which may be adjusted with field data. For precipitation data, international database can also be used in combination with local data if available. The real evapotranspiration (ETr) values may be estimated from the precipitation (P) data, using a empirical formula:
ETr: real evapotranspiration
Pavg: Average Precipitation data
ETravg: average evapotranspiration
P: daily precipitation data
We made 380 depth measurement points along one profile over the Kinkony lake. One flow measurement was done on Mahavavy river. A household surveys were conducted on a sample of the population; the sample size is 320. The questionnaire concerns their use of natural resources including water, woods and timber and medicinal plants.
Water account is established for year 2013 and for year 2018.
Maps in Fig 5 show the observed changes in water between 2013 and 2018, The next Figure shows the maps of water changes between 2013 and 2018: stocks, accessible resources, total water use, use index and the water ecological internal unit value.
Average Precipitation data is from WorldClim database [34], the average evapotranspiration is from CGIAR database [35] and the daily precipitation data are from CHIRPS database [36].
2.4 The ecosystem carbon account
The ecosystem carbon accounts focus on terrestrial and aquatic ecosystems with biomass and soil organic carbon stocks and flows. Fossil carbon stocks are not part of this account. Emissions of greenhouse gas from biomass burning or processes generating methane and volatile organic compounds (VOCs) are already recorded in the basic balances.
This Ecosystem carbon account records the stocks and flows of organic carbon available in soil, in underground and above-ground vegetation, and in water (fish and plant species). The gross primary production (GPP) flow of biomass by natural and cultivated vegetation is considered as a carbon flux.
The main data used in this account are the (i) above ground biomass that is obtained using data from local and/or regional ecological surveys and allometric equation [37] and the land cover maps. Since no data on underground carbon is generally available, underground carbon is assumed to show no change; (ii) Soil organic carbon values are, for example, from ISRIC Soil Grids with 250 m ground resolution [38]; (iii) GPP may be taken from MODIS website [39]; and (iv) household surveys provide information on quantity of resources used by local population. The Ecosystem Infrastructure functional services account.
2.5 The ecosystem landscape infrastructure functional services account
Infrastructure functional services are also called intangible services. ENCA assumes that an ecosystem in good condition continuously provides ecosystem services.
Therefore, ENCA uses landscape indicators to assess the potential of ecosystems to provide intangible ecosystem services.
Three fundamentals’ landscapes indices are used:
- The Green Landscape Background Index (GBLI) which represents the Land Cover biomass potential;
- The High Natural Value Index (HNVI) takes into account the level of protection of an area;
- The Landscape Fragmentation Index (FI) shows the level of fragmentation of the ecosystem. It is considered a negative effect.
The fragmentation index as proposed by Werber (2014) [11] is obtained using the Effective Mesh Size [40]. But 2 more indices are added in our study because in general there are very few roads and many paths. Inspired from McGarigal (1995) [41], the two new indices, the numbers of patches and surface ratio between tree cover and the river basin area, are added. To obtained a final fragmentation index, the average of these three indices are calculated.
The multiplication of these three indices (GBLI, HNVI, FI) gives the Landscape Ecosystem Potential or LEP [18].
For rivers, three other indices are also calculated:
- The River Accessibility Weighted Index is a measurement of a potential based on river extent (their length) and discharge (for each river section, length x log (discharge value));
- HNVI-river which is the result of intersection of HNVI with the rasterized rivers;
- River fragmentation which is the number of dams in each watershed.
The composite index of river potential is obtained by multiplying the last three indices. The index is then superposed with SELUs map. One obtains the mean index value for each SELU. This is then multiplied by the area of each SELU to obtain the River Ecosystem Potential (REP).
The Total Ecosystem Potential (TEIP) of the ecosystem infrastructure is the sum of the LEP and REP values.
2.6 The ecosystem capital capability account
The three ecosystem accounts each generate indices representing the unit ecological value of each ecosystem type. This is the result of multiplying the usage index by the ecosystem health index. When these values are then multiplied by the corresponding available resources, these indices collectively determine the overall ecosystem capacity, known as the ecosystem capability of the area. [18]
The capability of an ecosystem means its overall potential to deliver all services in a sustainable way without reducing the potential for other services. The difference between the capability values at the closing and opening dates is called the ecological debt, if negative, or ecological credit if positive. Ecological debt reflects the degree of degradation of ecosystems and ecological credit reflects the degree of improvement of the ecosystem states.
2.7 Study area
The Mahavavy-Kinkony Complex is a protected area located between 15°46’ and 16°12’ South latitude and 45°27’ and 45°56’ East longitude in North-western of Madagascar (see Fig 1). It covers some 350,000 ha. The northern part of the site opens onto the Mozambique Channel. The elevation is varying from 0 to 150 m [42]. It is a category V protected area according to the International Union for Conservation of Nature (IUCN) classification. Its status as protected area is obtained in 2015 [43]. The protected area includes several villages and two small cities. The landscape is composed of various ecosystems: forests (dry forest and mangrove), savannah, lakes, marine and coastal ecosystems. Mahavavy-Kinkony Complex is a refuge for endemic and threatened terrestrial and aquatic species found in different natural habitats such as lakes, rivers, swamps, mangroves, and dry forests. At the mouth of the Mahavavy River, the Mahavavy Delta is the largest area of mangroves [42]. The mangrove extends over 80 km along the coast. The Kinkony lake is the second largest one in Madagascar, and a number of threatened waterbird species can also be 250 seen here [42].
The region’s climate is a dry tropical climate with two contrasting seasons, dry season from April to October and rainy season from November to March. The total of annual rainfall is 1554 mm, with a maximum in January (475.6mm) and a minimum in June (0.6mm). The average annual temperature is 26°C, with a minimum of 18°C in July and a maximum of 35°C in December [42].
For the application of ENCA to the Mahavavy-Kinkony Complex protected area, the years of 2013 and 2018 have been selected as opening and closing dates. The choice of these two dates is due to the fact that the protected area of Mahavavy-Kinkony Complex was established in 2015 [36]. This allows for a comparison of the state of the ecosystem before the protection was established and the state of ecosystems after the establishment of the protected area status.
3. Results
Each ecosystem accounts are full presented on this section and the maps of some indicators. All maps were produced based on research results and generated using QGIS.
3.1 The land cover account
Table 3 shows land cover classes in the Mahavavy-Kinkony Complex protected area, and how they were derived from the primary land cover classes as proposed by Weber (2014).
The following Table 3 presents the different entities present in the land use maps. This nomenclature is derived from the original nomenclature proposed by Weber (2014) and which is compatible with other nomenclatures.
The following Fig 2 shows the two land cover maps of the year 2013 and the year 2018.
(a) maps in 2013 and in (b) in 2018.
The next Fig 3 shows the map of land cover changes between 2013 and 2018. It is obtained by taking the difference between the two previous maps.
The following Table 4 presents the diachronic analysis of land use, spanning both 2013 and 2018.
3.2 Socio-ecological landscape units
The next Fig 4 shows the maps of the SELU in 2013 and in 2018.
(a) in 2013 and (b) in 2018.
Changes in DLT (dominant land cover type) alert in areas under fast, such as on the North-East and South-West of the maps Fig 4.
3.3 The ecosystem water account
The Fig 5 shows the changes map of all important indices in water account.
(a) in drainage (b) lakes and reservoirs (c) accessible water resource (d) total of uses (the positive value shows that the uses of water resources are increasing between 2013 and 2018), (e) water use index and (f) water internal value.
The Tables 5 and 6 shows the ecosystem water account of Mahavavy-Kinkony Complex protected area in 2013 and 2018.
3.4 The ecosystem carbon accounts
The ecosystem carbon account is calculated for year 2013 and for year 2018. The Fig 6 shows the maps of opening stocks of carbon, the flows of the ecosystem carbon, the changes of total use of carbon, and the sustainable use index for 2013 and for 2018.
(a) opening stock of carbon for 2013 and (b) for 2018, (c) the changes of carbon accessible resources, (d) the variation of total use of carbon (the positive values shows that the use of carbon resources is increasing in 2018 compared to 2013), the sustainable use index for (e) 2013 and (f) for 2018.
Tables 7 and 8 shows the ecosystem carbon account of mahavavy-Kinkony Complex protected area in 2013 and 2018.
3.5 The ecosystem infrastructure functional services account
The Fig 7 shows three indices in ecosystem infrastructure functional services.
Green Background Landscape Index in (a) 2013 and in (b) 2018, Fragmentation Index in (c) 2013and in (d) 2018, River Accessibility Weighted Index in (e) 2013 and in (f) 2018.
The Table 9 shows the ecosystem infrastructure functional services account of mahavavy-Kinkony Complex protected area between 2013 and 2018.
3.6 The ecosystem capability account
For this account, three maps are show to illustrate the total ecosystem capabilities (see Fig 8) of Mahavavy-Kinkony Complex protected area and the trend of Total Ecosystem Capability in 5 years.
(a) 2013 and (b) 2018, (c) the trend in five years.
The Tables 10 and 11 shows the ecosystem capability account of mahavavy-Kinkony Complex protected area in 2013 and in 2018. The Table 12 show the difference in ecosystem capability between 2013 and 2018.
4. Discussions
Analysing the results of ecosystem natural capital accounting makes it possible to assess changes in the quantity and health of ecosystems. the discussion aims to clarify the information and indices derived from the accounting process, highlighting their implications for environmental sustainability, policy decisions and natural resource management.
4.1 The land cover account
The land cover account is shown in Table 4. The area of villages has increased by 119 ha (+27%) between 2013 and 2018. This trend can be considered as a proxy for population growth as the huts generally have the similar dimensions. According to the United Nations Population Fund [44] annual population growth in Madagascar is + 2.8% (+14% in five years). For mahavavy-Kinkony Complex this rate has almost doubled. The existence of migration is therefore justified. mahavavy-Kinkony Complex managers estimate a migration rate of + 2.5% [42]. These last values give a population growth of + 26.5% in 5 years.
For rice cultivation, an increase of 2,540 hectares (+60%) was recorded over five years. Most of this conversion came from natural land cover (vegetation including forest). Before the migration phenomenon, the main economic activities of local population are fishing (Kinkony lake and at sea), and working in the sugar factory of Namakia city. The new migrants settle in the south-east of lake Kinkony and cultivate rice and food crops. These people certainly come from the neighbouring regions of Marovoay and Ambato-Boeny, other rice growing areas. These migrants are no longer true climate migrants [45], as rice cultivation requires financial investments and relatively long time before harvest.
The establishment of 1,109 hectares (+26%) of sugarcane plantations over five years is linked to the revival of the sugarcane factory in the small city of Namakia.
The pressure of human activities has reduced the raffia area by 180 hectares (-19%) in five years. Raffia tree only grows in area with water. Reduction of raffia area means decrease of water resource. One cause of this phenomena is the creation of rice fields Regarding cultivated fields, the population is converting savannah and forests cover into agricultural land, resulting in the creation of 8,692 hectares (+146%) from natural cover (Savannah and forest). This is similar to the trend in rice cultivation. Savannah areas have increased, with a total gain of 3,894 hectares (+6%). The creation of 18,129 hectares (line F_LF4) is due to the disappearance of forest, mangroves, and shrubs. The consumption (disappearance) of 13,807 hectares of savannah (line C_LF5) is attributed to forest and shrubland restoration efforts [42] Forest cover decreased by 3,528 hectares due to selective logging for construction purposes. For sparse dry forests, the loss of 10,884 hectares (-25%) is mainly caused by illegal charcoal production [42], primarily intended to the major city of Majunga. Additionally, the collection of medicinal plants, hunting, and sacred rituals are common practices among villagers. These activities lead to the creation of paths within forests, increasing accessibility and contributing to forest degradation. This explains the significant degradation of sparse forests, amounting to 12,855 hectares.
The net total change in dense, degraded, and stunted mangroves is -223 hectares. Generally, this reduction is due to selective wood cutting for construction. However, sedimentation also impacts wetland areas by altering soil structure [46]. The sedimentation phenomenon is evidenced by the increase of 721 hectares of bare soil and sand, which affects mangroves.
The destruction of Phragmites in Lake Kinkony (-882 hectares or -38%) is linked to the creation of rice fields in the Phragmites marsh areas and uncontrolled fires in the region. This threat is exacerbated by natural events such as cyclones, which disperse Phragmites away from the lake during the dry season. The plants are either burned or perish on land. Destruction of Phragmites leads to biodiversity loss and water quality degradation. It is the natural habitat of birdlife, and also provides a suitable area for fish [47,48].
4.2 The ecosystem water account
The general trend (Tables 5 and 6) indicates a decrease in available water over five years (-20%). For lakes and reservoirs (line W1_1), one cause of this decline is sedimentation, which was already noted by Randriamasimanana and Rabarimanana (2011) [46]. For rivers (line W1_2), the difference between 2013 and 2018 is linked to reduced flow rates, which in turn are connected to decreased of precipitation. Similar phenomenon is observed in the total water inflow, comprising precipitation and natural inflows from upstream watersheds.
However, in agriculture, water use has increased by about +43% compared to 2013 level. This is due to the creation of new agricultural sites. This has led to changes in some SELUs.
Total water outflow includes evapotranspiration, anthropogenic use, and natural outflow downstream. The latter is generally tied to flow rates, as it partly reflects runoff from the area. The reduction (-17%) in total water outflow between 2013 (2 514 690) and 2018 (2 067 567) was observed.
The values of sustainability use index are below 1 except for savannahs and forests. According to Fig 5E, a small improvement (+0.1) in the situation can be seen in the north of the CMK, where sugar cane is grown. Accessible water resources in the area have increased, hence this improvement (see Table 6 line W6). The situation has not therefore improved naturally, but by accumulating more water to compensate for the increased demand.
Water quality has remained relatively stable compared to the 2018 situation.
4.3 The ecosystem carbon accounts
The general trend indicates a decrease in carbon over five years. Each SELU (socio-ecological landscape unit) includes some degree of forest cover, even if it is not dominant. The exploitation of this forest cover explains the observed trend in each SELU. Although other types of land cover exist, carbon is more concentrated in trees. However, wood harvesting has increased over the five-year period (see Tables 7- and 8-line C3). The total values of carbon uses are 3,182,504 for 2013 and 4,817,599 for 2018. It reflects overexploitation of wood, as these figures are abnormally high compared to population growth. In line C3, the extraction within the agricultural landscape SELU nearly tripled over five years, driven by the conversion of natural land to arable land (see land cover account Table 4).
According to carbon account lines C3_11 and C3_199 (see Tables 7 and 8), the values indicate an increase in agricultural production. It remains insufficient because production methods are still rudimentary.
For four out of the six SELUs, uses of carbon resource (SCU) are sustainable, as the use index are still equal to 1, except for savannah, which is scored 0.77 in 2018. This is due to population growth leading to an increase in demand of Bismarck palm. However, forests and mangroves have scores of 0.83 and 0.53 respectively in 2013, and 0.57 and 0.49 in 2018. These values indicate overexploitation of forests and mangroves. The main activities contributing to this are charcoal production and logging for construction. These activities are quite widespread although illegal and despite restrictions.
4.4 The ecosystem infrastructure functional services account
For the Green Background Landscape Index, the forest and mangrove SELUs show the highest values (see Table 9 and Fig 7A and 7B). For river accessibility water index, the highest values are found in urban, agricultural, and wetland SELUs (see Fig 7E and 7F). However, due to declining of river flow rates (as shown in the water account – Tables 5 and 6), these values have also decreased.
Since the hard-core boundaries of the protected area have not changed, the HNVI indicator remains the same over five years. Regarding landscape fragmentation, urban and wetland SELUs are the most fragmented. Agricultural SELUs show a decrease of fragmentation index value. This is because agricultural area is more uniform. Fragmentation index values for forests and mangroves remain stable.
Overall, EISU indicator (Table 9) shows a downward trend, except for agricultural SELUs. This is due to the conversion of some SELUs (including forest) into agricultural ones,
For human footprint SELUs, EISU = 0.49. This decrease is linked to the reduction of river accessibility weighted index (Table 9). This situation can be a harbinger of water supply issue in urban areas.
4.5 The ecosystem capability account
The creation of agricultural areas has led to an increase in the production of ecosystem services in these zones (increase of + 61% of Total Ecosystem Capability – TEC – see Table 12). However, this increase is primarily related to provisioning services, while other types of services are not represented. This is evidenced by the low value GBLI (global background landscape index) of these agricultural areas compared to others (see Table 9).
Over five years, the protected area experienced a -5% reduction in its capacity to provide all ecosystem services (see Table 12). This decline corresponds to an ecological debt of -275,164 ECU. This reduction is not solely linked to any specific natural resources but can be attributed to the simultaneous decline of multiple resources, highlighting the interconnections between different types of natural resources. In addition to reforestation programmes, it is recommended to address other phenomena, such as the silting-up of water surfaces inside the protected area.
4.6 Limits of the study
While the landscape units approach used in this study provides a valuable framework for linking land cover dynamics to broader land systems science and archetype literature, it has inherent limitations that should be acknowledged. One key limitation arises from the methodology employed to categorize and assess landscape units based on the dominant land cover. The use of dominant land cover can obscure significant changes within the landscape, especially in cases where a shift in land cover occurs but remains within the same landscape unit.
This limitation becomes particularly important when considering ecological processes within these landscape units. Such transformations can affect biodiversity, ecosystem services, and land management strategies, but these changes may not be accurately captured if only the dominant land cover is considered. Thus, while this methodology is a good starting point for analyzing land cover dynamics on a large scale, it may oversimplify landscape changes and overlook some important consequences.
One limitation of this study is the selection of only two specific years, 2013 and 2018, as the opening and closing dates for the application of the ENCA to the Mahavavy-Kinkony Complex protected area. While these years provide a useful comparison of the ecosystem state before and after the establishment of the protected area in 2015, they may not fully capture the dynamics and ongoing changes within the ecosystem over time. Additionally, the absence of data prior to 2013 limits the ability to assess long-term trends and pre-existing conditions of the area before the protected status was even considered. This narrower timeframe may overlook potential shifts in the ecosystem that occurred before or immediately after the establishment of the protection status, which could provide a more comprehensive understanding of the ecological impacts of the protected area.
On the other hand, the overall accuracy of the maps, expressed as a percentage (see appendix), is an indicator of their reliability. For 2013, the overall accuracy of 90% indicates that the land cover maps correctly identify the cover categories in 90% of cases. For 2018, an overall accuracy of 91% is slightly higher than in 2013, and indicates a decline in the reliability of land cover maps. Although these levels of accuracy are high, they warn of the need to continue refining classification methods and data processing to achieve an even more accurate representation of land landscapes.
Also, In-situ measurement data, obtained directly in the field, provides detailed and accurate information that reflects the real and specific conditions at each site. However, these data have limitations, such as their geographical representativeness. They are often limited to areas where observations and sampling have been carried out. International data from global sources can cover vast geographical areas and provide comparisons between different regions and ecosystems. However, these data also have their limitations, such as spatial resolution. International data may be insufficient to identify fine detail and heterogeneity within local ecosystems.
This study is exclusively focused on the mahavavy-Kinkony Complex protected area without taking into consideration interactions with the surrounding buffer zones. These interactions can affect trophic relationships and food webs. Buffer zones can also play a key role in ecological resilience by providing refuges or alternative resources in case of disturbances. However, it should be noted that buffer zones are generally difficult to access.
5. Conclusion
Our study on ecosystem natural capital account of the mahavavy-Kinkony Complex’s protected area revealed how ecosystem is evolving. It allows managers to make well-adapted decisions.
Results of ENCA analysis can be now shown as maps which are more easily understandable. ENCA recommends the valuation of ecological debts and credits on the basis of restoration costs. Ecological balance-sheets would be then important tool for integrating biodiversity into financial risks assessments. They would complement the carbon balances presently in use by financial or environmental institutions for climate change with much needed comprehensive ecological vision. Financial institutions such as Banque de France [49].are working for such enlargement. Similar approach is the proposed by Vardon et al (2021) [50]. They proposed the use of benchmark for an environmental bank. The value of ENCA ecological debt/credit can be used for that.
Supporting information
S1 Appendix. The confusion matrix of 2013 classification result.
https://doi.org/10.1371/journal.pone.0321948.s001
(DOCX)
Acknowledgments
We would like to express our appreciation to “Sud Expert Plantes Développement Durable” SEP2D and the Critical Ecosystem Partnership Fund (CEPF) for providing financial supports for this research project. The Critical Ecosystem Partnership Fund is a joint initiative of l’Agence Francaise de Développement, Conservation International, the European Union, the Global Environement Facility, the Government of Japan and the World Bank. A fundamental goal is to ensure civil society engaged in biodiversity conservation. We extend our warm thanks to the reviewers. Their comments allowed us to significantly improve this manuscript.
References
- 1.
Simon N. Red data book. Volume 1. Mammalia, a compilation. Switzerland; 1966. Available from: https://portals.iucn.org/library/node/6495
- 2.
Vincent J, Simon N. Red data book. Volume 2. Aves. Switzerland; 1966. Available from: https://portals.iucn.org/library/node/10048
- 3.
Meadows DL, Meadows DL, Randers J, Behrens WW III. The limits to growth: a report for the club of rome’s project on the predicament of mankind. New York: Universe Books; 1972. Available from: https://www.donellameadows.org/wp-content/userfiles/Limits-to-Growth-digital-scan-version.pdf
- 4.
United Nations. Report of the United Nations conference on the human environment, Stockholm 5-16 June 1972. New York: United Nations; 1973. Available from: https://documents-dds-ny.un.org/doc/UNDOC/GEN/NL7/300/05/PDF/NL730005.pdf?OpenElement
- 5. Brundtland G. Report of the world commission on environment and development: our common future. United Nations General Assembly document A/42/427; 1987. 400 p. Available from: http://www.un-documents.net/ocf-ov.htm
- 6. United Nations. United Nations Conference on Environment and Development; 1992 Jun 3–14; Rio de Janeiro, Brazil. United Nations; [cited 2022 Nov 8. ]. Available from: https://www.un.org/en/conferences/environment/rio1992
- 7. United Nations. AGENDA 21. United Nations Conference on Environment & Development; 1992 Jun 3–14; Rio de Janerio, Brazil [Internet]; 1992 [cited 2021 May 16. ]. Available from: https://sustainabledevelopment.un.org/content/documents/Agenda21.pdf
- 8. Ehrlich PR, Mooney HA. Extinction, substitution, and ecosystem services. BioScience. 1983;33(4):248–54.
- 9.
United Nations. System of environmental-economic accounting 2012 - experimental ecosystem accounting. New York; 2012. Available from: https://seea.un.org/sites/seea.un.org/files/seea_eea_final_en_1.pdf
- 10.
Leadley PW, Alkemade R, Pereira HM, Sumaila UR, Walpole M, Marques A, et al. Progress towards the aichi biodiversity targets: an assessment of biodiversity trends, policy scenarios and key actions. Montreal, CANADA; 2014. 500 p. Available from: https://www.cbd.int/doc/publications/cbd-ts-78-en.pdf
- 11.
Weber J-L. Ecosystem natural capital accounts: a quick start package. Technical Series. Montreal: Secretariat of the Convention on Biological Diversity; 2014.
- 12. United Nations. System of Environmental-Economic Accounting— Ecosystem Accounting (SEEA EA). White cover publication, pre-edited text subject to official editing; 2021. Available from: https://seea.un.org/sites/seea.un.org/files/documents/EA/seea_ea_white_cover_final.pdf
- 13.
World Bank. Guidebook on ecosystem accounting. Washington (DC); 2017. https://doi.org/10.1596/29829
- 14.
World Bank. Global Program on Sustainability (GPS) - annual report FY22. The World Bank; 2022. 90 p. Available from: https://thedocs.worldbank.org/en/doc/d3294f3f7a7aefa874c4cd794d098711-0320052023/original/global-program-on-sustainability-FY22-annual-report.pdf
- 15. Villa F, Bagstad KJ, Voigt B, Johnson GW, Portela R, Honzák M, et al. A methodology for adaptable and robust ecosystem services assessment. PLoS One. 2014;9(3):e91001. pmid:24625496
- 16. Rambaud A, Richard J. The “Triple Depreciation Line” instead of the “Triple Bottom Line”: towards a genuine integrated reporting. Crit Perspect Account. 2015;33:92–116.
- 17.
Natural Capital Project. InVEST 3.12.1. post6+ug. g3c8454b User’s Guide. Stanford University, University of Minnesota, Chinese Academy of Sciences, The Nature Conservancy, World Wildlife Fund, and Stockholm Resilience Centre; 2022. Available from: https://storage.googleapis.com/releases.naturalcapitalproject.org/invest-userguide/latest/index.html
- 18.
Rahm M, Lardeux C, Weber J-L, Ramihangihajason TA. Ecosystem natural capital accounting in the Guiana Shield in 2000 and 2015. Application of the quick start package of the convention on biological diversity (CBD) for implementing Aichi biodiversity target 2 on integration of biodiversity values in national accounting systems. ECOSEO Project. WWF; 2020.
- 19.
Weber J-L, Tapsoba T, Mar FN, Romdhane A, Tafi J, Fourmann E, et al. ArfikENCA 2001-2020 Une première série de comptes écosystémiques pour le continent africain. Observatoire du Sahara et du Sahel; 2023. 34 p. Available from: https://oss-online.org/sites/default/files/2023-05/ArfikENCA.pdf
- 20. Morand J-A, Jaffrain G, Sannier C, Weber J-L. Experimental ecosystem natural capital accounts in the Republic of Guinea. Int Arch Photogramm Remote Sens Spatial Inf Sci. 2023;XLVIII-M-1–2023:203–10.
- 21.
TEEB, editor. Mainstreaming the economics of nature: a synthesis of the approach, conclusions and recommendations of teeb. Geneva: UNEP; 2010.
- 22. Buchhorn M, Van den Hoof C, Smets B, Weber J-L, Sanon AA, Tiemtoré S. Facilitating the management of protected areas through multi-level ecosystem accounting on an example in West Africa. Sustainability. 2023;15(12):9198.
- 23.
Di Gregorio A, Weber J-L, Jaffrain G. Land cover classification for ecosystem accounting. ISSUE 3: Land cover mapping, land cover classifications, and accounting units. London UK; 2011. Available from: https://unstats.un.org/unsd/envaccounting/seeales/egm/Issue3_EEA_FAO.pdf
- 24.
Heymann Y, Steenmans C, Croiselle G, Bossard M. CORINE land cover technical guide. Brussels, Luxembourg: European Commission, CECA-CEE-CEEA; 1993.
- 25.
Bossard M, Feranec J, Oťaheľ J. Corine land cover—technical guide. Copenhagen, Denmark: European Environment Agency; 2000.
- 26.
David AK, Ferrer-Paris JR, Nicholson E, Kingsford RT. IUCN global ecosystem typology 2.0: descriptive profiles for biomes and ecosystem functional groups. Gland, Switzerland: IUCN, International Union for Conservation of Nature; 2020.
- 27.
USGS. Landsat 8 | U.S. Geological Survey. In: U.S. Geological Survey [Internet]; 2010 [cited 2024 Nov 17. ]. Available from: https://www.usgs.gov/landsat-missions/landsat-8
- 28. Sun W, Chen B, Messinger DW. Nearest-neighbor diffusion-based pan-sharpening algorithm for spectral images. Opt Eng. 2014;53(1):013107.
- 29. Gislason PO, Benediktsson JA, Sveinsson JR. Random forests for land cover classification. Pattern Recognit Lett. 2006;27(4):294–300.
- 30. Olofsson P, Foody GM, Herold M, Stehman SV, Woodcock CE, Wulder MA. Good practices for estimating area and assessing accuracy of land change. Remote Sens Environ. 2014;148:42–57.
- 31. Serra P, Pons X, Saurí D. Post-classification change detection with data from different sensors: some accuracy considerations. Int J Remote Sens. 2003;24(16):3311–40.
- 32. Jensen JR. Introductory digital image processing: a remote sensing perspective.; 1996 [cited 2023 Jan 30. ]. Available from: https://www.cabdirect.org/cabdirect/abstract/20001911540
- 33. Johary ARF, Rakotondraompiana S, Ravonjimalala R, Rakotoniaina S. Détection des changements d’occupation du sol avec la méthode IR-MAD. Application à la forêt sèche des Mikea (S.-O. de Madagascar). 2018;2:24–32.
- 34. Fick SE, Hijmans RJ. WorldClim 2: new 1‐km spatial resolution climate surfaces for global land areas. Int J Climatol. 2017;37(12):4302–15.
- 35. Trabucco A, Zomer RJ. Global aridity index and potential evapo-transpiration (ET0) climate database v2; 2019;10.
- 36. CHIRPS. Rainfall estimates from rain gauge and satellite observations | Climate Hazards Center - UC Santa Barbara. [cited 2024 Nov 18. ]. Available from: https://www.chc.ucsb.edu/data/chirps
- 37. Chave J, Réjou-Méchain M, Búrquez A, Chidumayo E, Colgan MS, Delitti WBC, et al. Improved allometric models to estimate the aboveground biomass of tropical trees. Glob Chang Biol. 2014;20(10):3177–90. pmid:24817483
- 38. Hengl T, Mendes de Jesus J, Heuvelink GBM, Ruiperez Gonzalez M, Kilibarda M, Blagotić A, et al. SoilGrids250m: Global gridded soil information based on machine learning. PLoS One. 2017;12(2):e0169748. pmid:28207752
- 39. Endsley KA, Zhao M, Kimball JS, Devadiga S. Continuity of global MODIS terrestrial primary productivity estimates in the VIIRS era using model‐data fusion. JGR Biogeosci. 2023;128(9).
- 40. Jaeger JAG. Landscape division, splitting index, and effective mesh size: new measures of landscape fragmentation. Landsc Ecol. 2000;15(2):115–30.
- 41.
McGarigal K. FRAGSTATS: spatial pattern analysis program for quantifying landscape structure. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station; 1995.
- 42.
Asity Madagascar. Plan d’aménagement et de gestion du Complexe des zones humides de Mahavavy-Kinkony, Région Boeny. Antananarivo; 2015.
- 43.
Biodiversity Informatics Platform. Protected Areas. In: Protected Areas [Internet]; 2020 [cited 2024 Nov 17. ]. Available from: https://protectedareas.mg/landscape/show/8
- 44.
UNFPA. Population et développement. In: UNFPA Madagascar [Internet]; 2016 Feb 25 [cited 2023 Nov 23. ]. Available from: https://madagascar.unfpa.org/fr/topics/population-et-d%C3%A9veloppement
- 45.
Harioly Nirina MO. Les enjeux de la Migration « liée à la sécheresse » dans la Région ANDROY, Sud de MADAGASCAR. Liège, Belgique: Université de Liège; 2018. Available from: https://matheo.uliege.be/handle/2268.2/5567
- 46. Andriamasimanana R. Analyses de la dégradation du lac Kinkony pour la conservation du Complexe des Zones Humides Mahavavy - Kinkony, Région Boeny, Madagascar. MCD. 2011;6(1).
- 47. Whyte RS, Bocetti CI, Klarer DM. Bird assemblages inphragmitesdominated and non-phragmiteshabitats in two lake erie coastal marshes. Nat Areas J. 2015;35(2):235–45.
- 48.
Wynia A. Fish and invertebrate use of invasive Phragmites in a Great Lakes freshwater delta; 2019. https://doi.org/10.13140/RG.2.2.24829.51681
- 49.
Salin M, Svartzaman R, Espagne E, Gauthey J, Hadji-Lazaro P, Allen T, et al. Perte de biodiversité et stabilité financière. In: Banque de France [Internet]; 2022 [cited 2024 Jul 24. ]. Available from: https://www.banque-france.fr/fr/publications-et-statistiques/publications/perte-de-biodiversite-et-stabilite-financiere
- 50. Vardon MJ, Keith H, Burnett P, Lindenmayer DB. From natural capital accounting to natural capital banking. Nat Sustain. 2021;4(10):832–4.