Case Study: Madagascar
1. Outcome
This Circular documents Madagascar's first SEEA Ecosystem Accounting condition account for photic coral reefs (M1.3), compiled for 27 sites along the Atsimo-Andrefana coast (southwest Madagascar) over two survey campaigns—November 2025 (opening period) and January–February 2026 (closing period). It demonstrates how standardised underwater visual census, allometric biomass estimation and Western Indian Ocean (WIO) reference levels can be combined into a repeatable, multi-indicator condition account that is ready for annual update and that provides the biological foundation for a subsequent ecosystem-services valuation.
2. Requirements
3. Case Study Material
3.1 Country Context
The Atsimo-Andrefana region of southwest Madagascar fronts one of the most extensive fringing- and barrier-reef systems in the Western Indian Ocean. The reefs underpin small-scale fisheries, coastal tourism and shoreline protection for Vezo fishing communities, and are managed through a network of Locally Managed Marine Areas (LMMAs) and customary governance arrangements (dina) operating alongside national fisheries regulation. Despite the system's regional importance, no standardised, repeatable measure of reef condition existed prior to this pilot, and no reef-system value has previously been recorded in national budget or NDC reporting.
3.2 Implementation Approach
The pilot combined two campaigns of underwater visual census (UVC) with WIO regional reference levels and a SEEA-EA condition index (CI) on a 0–1 scale:
| Component | Periods | Method |
|---|---|---|
| Coral-reef fish (M1.3) | Nov 2025 (opening); Jan–Feb 2026 (closing) | UVC along 2 transects × 85 stations × 27 sites; 6 size-class bins; allometric W = a · L^b biomass estimation; 5,186 fish records, 227 species, 39 families |
| Reef macroinvertebrates | Nov 2025–Feb 2026 | Abundance counts on 3 transects × 81 stations; 59 species (molluscs, echinoderms, crustaceans, holothurians); 434 records post-QA/QC |
| Sea urchins | Nov 2025–Feb 2026 | Abundance counts on 2 transects × 79 stations; 8 species; 264 records |
| Crown-of-thorns starfish (COTS) | Nov 2025–Feb 2026 | Site-level density; outbreak threshold 15 ind/ha (Dulvy et al. 2021) |
Opening- and closing-period site groups are partly disjoint (15 sites surveyed in 2025; 12 sites surveyed in 2026; see §3.6); cross-period comparison therefore reflects spatial variation among site groups rather than temporal change at repeated sites.
3.3 Worked Accounts
3.3.1 SEEA-EA condition account—coral reef (M1.3)
Region-level indicator account, both survey periods, against published WIO reference levels:
| Year | Indicator | Reference level | Measured value | Condition index |
|---|---|---|---|---|
| 2025 | Fish biomass | 1,150 kg/ha (WIO conservation target; McClanahan et al. 2016) | 771.6 kg/ha | 0.58 |
| 2025 | Fish species richness | 94 species (max observed, 2025) | 55 species | 0.59 |
| 2025 | COTS density (inverted) | 0 ind/ha (ideal) | 2.7 ind/ha | 0.84 |
| 2026 | Fish biomass | 1,150 kg/ha | 710.1 kg/ha | 0.59 |
| 2026 | Fish species richness | 103 species (max observed, 2026) | 56 species | 0.54 |
| 2026 | COTS density (inverted) | 0 ind/ha (ideal) | 0.0 ind/ha | 1.00 |
Region-level means mask substantial site-to-site variation: fish biomass at site level ranges from 499 kg/ha (Ambatomilo) to 4,707 kg/ha (highest-biomass sites), and species richness from 25 to 103 species per site.
3.3.2 Fish biomass—selected sites (CI 0–1, normalised to 1,150 kg/ha)
| Site | Period | CI (fish biomass) |
|---|---|---|
| Ankaramifioky | 2026 | 1.00 |
| Salary | 2026 | 1.00 |
| Tsandamba | 2026 | 0.98 |
| Ambitiky | 2025 | 1.00 |
| Beheloky | 2025 | 1.00 |
| Retsela | 2025 | 1.00 |
| Ambohibola | 2025 | 0.88 |
| Itampolo | 2025 | 0.74 |
| Belavenoky | 2026 | 0.82 |
| Ambatomilo | 2026 | 0.17 |
| Ifaty | 2025 | 0.21 |
| Fitsitika | 2025 | 0.22 |
| Ampasindava | 2026 | 0.21 |
Six sites are at or above the 1,150 kg/ha WIO managed-reef target (CI = 1.00); two sites (Ambatomilo, Ampasindava) sit below 20% of target.
3.3.3 Species richness—selected sites (CI normalised to within-period maximum)
| Site | Period | CI (richness) |
|---|---|---|
| Mangily | 2025 | 1.00 |
| Belavenoky | 2026 | 1.00 |
| Andavadoaky | 2026 | 0.97 |
| Salary | 2026 | 0.83 |
| Beheloky | 2025 | 0.82 |
| Ambola | 2025 | 0.74 |
| Retsela | 2025 | 0.74 |
| Ifaty / Itampolo | 2025 | 0.72 |
| Ambatomilo | 2026 | 0.33 |
| Ampasindava | 2026 | 0.24 |
The aggregate richness CI (0.54–0.59) sits well below the fish-biomass CI—the characteristic signature of selective fishing, in which large slow-reproducing species are removed first, leaving smaller fast-reproducing taxa to maintain biomass.
3.3.4 COTS outbreaks (2025 opening period)
| Site | COTS density (ind/ha) | Status |
|---|---|---|
| Anakao | 25 | Outbreak (>15 ind/ha) |
| Beheloky | 25 | Outbreak |
| Retsela | 50 | Outbreak |
| All other 24 sites | 0 | At reference |
Three of 27 sites are in active outbreak condition against the 15 ind/ha threshold (Dulvy et al. 2021); healthy WIO reefs with intact predatory-fish layers typically record < 1 ind/ha.
3.3.5 Complementary invertebrate indicators
| Indicator | Mean (range across sites) | Provisional reference | Note |
|---|---|---|---|
| Sea urchin density | 0.7 ind/m² (0 – 2.9) | 1.5–2.0 ind/m² (Echinometra mathaei, WIO; LOW confidence) | Below provisional reference; consistent with predator-mediated control |
| Giant clam abundance | 1.6 per transect (0 – 16.5) | -- | High variability; site-specific |
| Macroinvertebrate richness | 8.6 taxa per site (2 – 20) | -- | Used as complementary diagnostic, not currently normalised to CI |
3.4 Decision Applications
- Early-warning signal for fisheries management. High biomass (CI 0.58–0.59) combined with depressed diversity (CI 0.54–0.59) is a documented precursor to fishery productivity loss. The account identifies the inflection point while management options are still preventive rather than remedial.
- Targeted enforcement. Site-level CIs identify three COTS-outbreak sites (Anakao, Beheloky, Retsela) and two low-biomass sites (Ambatomilo, Ampasindava) for priority enforcement and management response.
- NDC and GBF reporting. Condition indices directly support Madagascar's 2023 NDC Section 3.2 (Ecosystem-Based Adaptation) annual ecosystem-resilience monitoring commitment, and supply indicators for Kunming–Montreal Global Biodiversity Framework Target 8.
- Foundation for valuation. Fish-biomass data feed a planned ecosystem-services valuation (wild fish provisioning, nursery function, reef-based tourism), pending confirmed reef extent, local market prices and MPRH landings—targeted for mid-2026.
- Regional positioning. Madagascar is the first WIO state with a SEEA-EA-format reef condition baseline at this resolution; the protocol is directly replicable across the region.
3.5 Lessons Learned
- Within-period maximum is a defensible interim reference for richness. Published WIO reference ranges (82–152 species per site) are too sparse and habitat-heterogeneous for site-level normalisation; using the observed maximum per period preserves comparability across sites within a year and is transparently re-baselined annually.
- Aggregate condition masks management-relevant variation. Region-mean fish-biomass CI (0.58–0.59) conceals a tenfold spread at site level (CI 0.17 → 1.00). Site-level disaggregation is essential for any management or enforcement application.
- Disjoint opening/closing site groups limit trend inference. Until the same 27 sites are re-surveyed, cross-period change cannot be interpreted as temporal trend. A repeat survey by mid-2026 is the binding pre-requisite for a defensible 2025–2026 change account.
- Multi-indicator framing strengthens diagnostic power. Fish biomass alone would have reported a healthy reef; pairing biomass with richness and COTS density reveals selective-fishing pressure and localised outbreaks that single-indicator accounts would miss.
- Sensitivity to reference levels. Sea urchin and (where used) macroinvertebrate references are LOW-confidence WIO provisional values; CI movements should be interpreted alongside the underlying measured value, not in isolation.
3.6 Data Gaps and Caveats
- Opening (2025; 15 sites) and closing (2026; 12 sites) site groups are partly disjoint; CI change cannot yet be attributed to temporal trend.
- Fish allometric a/b parameters: 124 unique a values and 55 unique b values applied; 50–60 cm size-class is sparsely populated (one unique record across the dataset).
- Transect-area assumption pending review (see source plan §1).
- Sea-urchin reference level is provisional WIO (LOW confidence); macroinvertebrate richness is not yet normalised to a CI.
- Ecosystem-services valuation (fish provisioning, nursery function, tourism) is not yet compiled—pending confirmed reef extent (satellite mapping, 2026), local market-price survey, and annual landings data from Madagascar MPRH.
- No total economic value or per-household income figure is claimed in this account.
4. Acknowledgements
Authors: [To be confirmed]
Reviewers: [To be confirmed]
Data partners: Madagascar Ministry of Fisheries and Blue Economy (MPRH); Atsimo-Andrefana regional fisheries authorities; LMMA and dina community governance bodies; field observers Laza (fish) and Maka (invertebrates).
5. References
Source data: Madagascar_accounts/03_outputs/ (MDG_fish_invert_seea_condition.csv, MDG_fish_invert_condition_account_normalized.csv, MDG_fish_invert_site_condition.csv, MDG_fish_invert_site_fg_condition.csv). Source documentation: Madagascar_accounts/docs/accounts/cond_reef_fish_invert_plan.md; Madagascar_accounts/docs/briefs/POLICY_BRIEF_Madagascar_CoralReef.md. Reference benchmarks: McClanahan et al. (2016); Dulvy et al. (2021); GCRMN (2021); Obura et al. (2022).