Freshwater biodiversity is not adequately addressed by the current protected areas of the Caribbean biodiversity hotspot

This study reveals that current protected areas in Cuba inadequately cover freshwater biodiversity, necessitating a significant expansion of conservation efforts—particularly in headwaters and for endemic species—to meet global targets, while suggesting that excluding existing reserves in planning could optimize resource efficiency.

The Gist

Imagine the Caribbean islands, specifically Cuba, as a giant, vibrant aquarium. Inside this aquarium, there are thousands of unique fish, tiny insects, and water plants that you won't find anywhere else on Earth. These creatures are the "special guests" of the island's freshwater ecosystems.

Now, imagine the government has built a series of "fenced-off zones" (Protected Areas) to keep these guests safe. The goal is to make sure at least 17% (and soon 30%) of every species' home is inside these safe zones.

The Problem: The Fences Are in the Wrong Place
This study is like a detective checking the map of those fences. The researchers found a big problem: The fences are mostly protecting the "living room" of the river (the bottom, near the ocean), but they are ignoring the "kitchen" and "bedroom" (the headwaters and upper streams).

Think of a river like a tree. The trunk is the main river near the sea, and the branches are the small streams high up in the mountains. Currently, Cuba's protected areas are mostly guarding the thick trunk. But many of the unique species live on the tiny, fragile branches high up. Because the fences don't reach the branches, these species are still in danger, even though they are technically "protected."

The Investigation: How They Found the Gaps
The researchers used a high-tech "digital crystal ball" (computer models) to predict exactly where 227 different species of water bugs, fish, and plants live. They then overlaid a map of the current protected areas onto these predictions.

• The Result: It was a disaster for many species. About 41% of the species didn't even have 17% of their home protected. When they looked at the stricter goal of 30% (the new global target), a whopping 71% of the species failed the test.
• The Victims: The most vulnerable were the "endemics"—the species that live only in Cuba. If their small patch of river isn't protected, they go extinct forever. Some famous endangered species, like the American Eel and a blind cavefish, are essentially unprotected.

The Solution: Two Ways to Fix the Map
The researchers ran two different simulations to see how to fix this:

1. The "Lock-In" Strategy (The Conservative Approach):
   Imagine you say, "We must keep all the fences we already built, but we need to add more to make it work."

   • The Catch: To make this work, you have to protect a massive amount of extra land. It's like trying to fill a bucket that already has holes in it; you need to pour in a lot more water just to get it to the same level. This approach is expensive and requires protecting 1.7 times more area than the current system.

2. The "Free-Choice" Strategy (The Smart Approach):
   Imagine you say, "Let's forget the old fences for a moment. Let's draw the perfect new map from scratch to catch all the species."

   • The Result: This approach found that you could protect the same number of species using less land and less money. It turns out that about 63% of the land currently designated as "protected" isn't actually needed to save these specific water creatures. The current system is inefficient; it's like guarding a castle with a moat that's too wide in the wrong places and too narrow in the right ones.

The Big Picture: Why This Matters
The study highlights that nature doesn't care about political borders or where we decided to put fences 20 years ago. Rivers flow from the mountains to the sea. If you protect the bottom but not the top, the water quality at the bottom gets ruined, and the creatures at the top die out.

The Takeaway for Everyone
To save Cuba's unique freshwater life, we need to:

• Look Upstream: Move the protection zones to the headwaters (the top of the river tree).
• Connect the Dots: Make sure the protected areas are linked together so animals can swim freely from the mountains to the sea.
• Be Smart with Money: Don't just keep the old, inefficient fences. Redraw the map to protect the most critical areas first.

In short, the current safety net has too many holes in the wrong places. To truly save the "hidden biodiversity" of Cuba's rivers, we need to weave a new, smarter net that covers the whole river, from the highest mountain spring to the ocean.

Technical Summary

1. Problem Statement

Freshwater ecosystems are experiencing a global decline in biodiversity at a rate exceeding that of marine and terrestrial systems. While Protected Areas (PAs) are a primary conservation tool, their effectiveness in safeguarding freshwater biodiversity remains uncertain.

• Context: The study focuses on the Cuban archipelago, a critical part of the Caribbean biodiversity hotspot known for high species richness and endemism.
• The Gap: Despite Cuba's National System of Protected Areas (SNAP) covering ~16% of the national surface (near the former Aichi Target 11 of 17%), there is a lack of evidence regarding its representativeness for freshwater species. Previous gap analyses were limited to specific taxa (e.g., fish, diving beetles) and did not account for the full spectrum of freshwater biodiversity or the post-2020 global biodiversity framework targets (30% protection).
• Specific Challenge: Current PA networks often fail to account for longitudinal connectivity (upstream-downstream flow) and lateral connectivity (river-floodplain interactions), which are crucial for aquatic species survival, especially under climate change.

2. Methodology

The authors employed a systematic spatial conservation prioritization approach using the following technical workflow:

A. Conservation Planning Units

• Unit Definition: Sub-catchments were used as planning units to integrate freshwater and terrestrial realms.
• Data Source: Derived from the MERIT Hydro Digital Elevation Model (3-arc sec resolution).
• Filtering: Initial 69,528 sub-catchments were filtered to 29,426 based on overlap with observed water bodies (streams, rivers, reservoirs, wetlands) from OpenStreetMap.

B. Species Data & Modeling

• Taxonomic Scope: 227 species total:
  • 182 Freshwater macroinvertebrates (macroinvertebrates).
  • 26 Vertebrates (primarily fish).
  • 19 Macrophytes.
• Data Processing: Occurrence records were filtered for georeferencing errors, temporal relevance (1970–present), and coordinate uncertainty (<5km). Only species with $\ge$10 records were retained.
• Pseudo-absence Generation: A two-step method (One-class SVM + random selection) was used to generate pseudo-absences with a fixed prevalence of 0.1.
• Ensemble Species Distribution Modeling (SDM): Four algorithms were combined via a committee average ensemble:
  1. Maxent (Maximum Entropy).
  2. BRT (Boosted Regression Trees).
  3. RF (Random Forest).
  4. SSN (Spatial Stream Networks) – specifically to handle spatial autocorrelation in dendritic networks (using tail-up, tail-down, and Euclidean distance models based on species dispersal behavior).
• Environmental Predictors: 10 principal components derived from climate, land cover, topography, soil, and river topology variables.
• Thresholding: Predictions were converted to binary presence/absence using a threshold minimizing the difference between sensitivity and specificity (AUC $\ge$ 0.7, TSS $\ge$ 0.7).

C. Spatial Conservation Prioritization

• Objective: Identify a minimum set of sub-catchments to protect 30% of the suitable habitat for all 227 species (aligning with the Kunming-Montreal Global Biodiversity Framework).
• Cost Metric: The Human Footprint Index (HFI) (1995–2004) was used as a cost layer; lower HFI (less human impact) = lower cost.
• Connectivity Penalty: A penalty term based on inverse distance between planning units within the same drainage basin was added to promote longitudinal connectivity.
• Two Scenarios:
  1. Lock-in-PA: The existing SNAP was "locked in" (mandatory inclusion), and the algorithm added necessary sub-catchments to meet targets.
  2. Free-choice-PA: The algorithm selected sub-catchments freely without being constrained by existing SNAP boundaries.
• Solver: Exact integer linear programming (Gurobi) via the R package prioritizr.

3. Key Results

A. Gap Analysis (Current Protection Status)

• Overall Failure: 41% of species (90 species) failed to meet the 17% Aichi target, and 71% (161 species) failed to meet the 30% post-2020 target under the current SNAP.
• Taxonomic Disparities:
  • Vertebrates (Fish): Received the lowest relative protection on average.
  • Macrophytes & Odonates: Better represented but still insufficient for the 30% target.
  • Threatened Species: Four species (including the Endangered damselfly Protoneura caligata and Vulnerable American Eel Anguilla rostrata) are critically underprotected.
• Endemism: At least 24 endemic species currently fail the 17% target; this number is projected to rise to 47 endemic species failing the 30% target by 2030.

B. Conservation Planning Solutions

• Lock-in-PA Solution:
  • Required protecting 19,437 km² (12,388 planning units).
  • This area is 1.7 times larger than the current inland water area of the SNAP.
  • Median Coverage: 45.21% across species.
  • Cost: Higher (2051.0 cost units) due to redundancy in existing protected areas.
• Free-choice-PA Solution:
  • Required protecting 14,229 km² (9,493 planning units).
  • Only 37% of this solution overlaps with the current SNAP, indicating significant inefficiency in the current network layout.
  • Cost: Lower (1524.8 cost units).
• Connectivity & Location:
  • Current SNAP tends to protect areas near river mouths, neglecting upstream headwaters.
  • Both optimization solutions successfully integrated headwaters and connected fragmented river segments, significantly improving longitudinal connectivity.
  • Priority Regions: Eastern Cuba (e.g., Cauto, Mayari, Toa basins) emerged as the most critical hotspots, followed by central and western regions.

4. Key Contributions

1. Comprehensive Gap Analysis: First study to assess the representativeness of the Cuban SNAP across a broad spectrum of freshwater taxa (macroinvertebrates, vertebrates, macrophytes) rather than just charismatic or single-species groups.
2. Methodological Integration: Successfully combined Spatial Stream Network (SSN) models with ensemble SDMs to account for the unique dendritic topology of river systems, improving distribution predictions for aquatic organisms.
3. Strategic Comparison: Provided a quantitative comparison between "locking in" existing PAs versus a "free-choice" approach, demonstrating that the current SNAP is spatially inefficient for freshwater conservation.
4. Policy Alignment: Directly addresses the Kunming-Montreal Global Biodiversity Framework (Target 3) by modeling the specific area and configuration required to meet the 30% inland water protection target.

5. Significance and Implications

• Urgency for Redesign: The study proves that simply expanding the current SNAP is insufficient; a strategic redesign is necessary. The current system protects terrestrial landscapes well but fails to capture the longitudinal connectivity required for freshwater species.
• Headwater Protection: The results highlight that upstream/headwater regions are currently neglected but are essential for meeting conservation targets. Future planning must prioritize these areas to ensure species persistence.
• Cost-Efficiency: The "free-choice" scenario suggests that achieving biodiversity goals could be done with fewer resources if the current PA network is re-evaluated and not blindly locked in, though a hybrid approach (Lock-in-PA) offers the most comprehensive connectivity.
• Threat Mitigation: By identifying specific gaps for endemic and threatened species (e.g., cavefish, endemic damselflies), the study provides a roadmap for preventing extinctions in the face of climate change, dam construction, and habitat fragmentation.
• Management Recommendations: The authors call for updating management plans to include freshwater species, strengthening monitoring of dam impacts, and engaging local communities to protect the identified priority basins in Eastern Cuba.

In conclusion, the paper provides robust evidence that the current protected area network in the Caribbean hotspot (Cuba) is inadequate for freshwater biodiversity. It offers a data-driven, spatially explicit roadmap to optimize conservation efforts, ensuring that the 30% global target is met efficiently while maintaining the ecological connectivity vital for riverine ecosystems.