Skip to content
Ecological Intelligence

Ecological evidence.
Independent perspective.
Explore Wild Lands ↗

Ecological Intelligence / Subject

spatial modeling

Peer-reviewed Europe

Tiny field-edge strips changed where pollinator models placed habitat

Evidence
Peer-reviewed
Geography
Three agricultural landscapes, Netherlands
System
Agricultural landscape / pollinator habitat

WHAT THE EVIDENCE SAYS

Fine-resolution (10-meter) species-distribution models incorporated herbaceous linear landscape elements as cover or distance covariates. Compared with models lacking these features, inclusion improved predictions and reduced overprediction in fields and underprediction at edges; improvements were evaluated against independently collected field records as well as withheld training data.

WHY IT MATTERS / OUR READ

Restoration opportunity maps need to capture narrow field margins and ditches, not just broad land-cover classes; independent field validation matters when models guide scarce conservation funds.

CAVEATS / UNCERTAINTY

Model improvement is not proof that newly adding a strip will cause equivalent abundance gains. Transferability from three Dutch landscapes to other crops, climates or taxa remains to be tested.

SOURCE

Landscape Ecology — Moens et al. (2026)
View original evidence ↗

Dutch bee and hoverfly models improved when linear habitat elements were mapped instead of being averaged into cropland.

Read the Wild Lands assessment →
Peer-reviewed Global

High flood-model accuracy can hide how a map will fail in the next flood

Evidence
Peer-reviewed
Geography
Global published flood-susceptibility modeling literature, 2010–2024
System
Flood-risk landscapes; GIS modeling; spatial validation

WHAT THE EVIDENCE SAYS

A review of 200 peer-reviewed flood-susceptibility modeling studies found that roughly 73% used validation that did not fully account for spatial dependence, likely exaggerating generalization. About 48% did not quantify uncertainty, and only around 6% used a probability-calibration assessment.

WHY IT MATTERS / OUR READ

The counter-signal is that an impressive AUC on a map is not a flood prediction people can safely use. Model evaluation should hold out spatially independent areas, expose residual uncertainty, and test the actual planning decision that a map is intended to support.

CAVEATS / UNCERTAINTY

The counts describe practices in a heterogeneous sample of published studies, not a meta-analytic correction for every model. The estimated validation inflation varies by data and design; the review does not establish accuracy for a specific floodplain.

SOURCE

Frontiers in Environmental Science · AI flood-mapping methodological review
View original evidence ↗

A 200-study review found widespread spatially naive validation and thin uncertainty reporting in AI susceptibility maps.

Read the Wild Lands assessment →
Europe

Baltic ecosystem risk concentrates where species hotspots overlap oxygen depletion and fishing pressure

Evidence
Peer-reviewed data-driven ecosystem risk modeling
Geography
Baltic Sea
System
Marine / Baltic Sea ecosystem

WHAT THE EVIDENCE SAYS

The study integrated 2020 distributions for 145 marine species with environmental pressures and used unsupervised clustering and representation learning to identify ecological-risk patterns. Species-abundant hotspots were especially exposed where bottom-oxygen depletion and fishing pressure overlapped.

WHY IT MATTERS / OUR READ

Multi-stressor maps are most useful when they show where ecological value and pressure coincide, not merely where either is high. That can sharpen monitoring and spatial-management priorities before one pressure is incorrectly treated as the whole problem.

CAVEATS / UNCERTAINTY

This is a model-based assessment built from a 2020 multi-source snapshot. Clusters indicate co-occurrence and relative risk, not causal attribution or measured future population loss, and results depend on the input layers and modeling choices.

SOURCE

Scientific Reports
View original evidence ↗

An unsupervised multi-stressor analysis of 2020 Baltic data identified species-rich areas where low bottom oxygen and fishing pressure combine into elevated ecological risk.

Read the Wild Lands assessment →
Peer-reviewed

Seasonal livestock maps can separate where grazing pressure actually occurs

Evidence
Peer-reviewed
Geography
China grazing-production systems
System
Rangelands / pastoral systems

WHAT THE EVIDENCE SAYS

A new dataset reconstructs 1-km warm- and cold-season distributions of cattle and sheep/goats across China from 2000–2021 using a stacking ensemble. Internal test performance was high and external city-level validation was lower but still useful, providing a spatially explicit alternative to annual aggregate livestock counts.

WHY IT MATTERS / OUR READ

Seasonal maps can make grazing exposure more ecologically interpretable than administrative totals, especially where animals move between summer and winter ranges.

CAVEATS / UNCERTAINTY

These are modeled distributions, not direct animal counts in every grid cell. External validation is weaker than internal performance, and livestock density alone does not establish ecological impact without vegetation, climate and management context.

SOURCE

Scientific Data
View original evidence ↗

EvidencePeer-reviewedGeographyChina grazing-production systemsSystemRangelands / pastoral systems WHAT THE EVIDENCE SAYS A new dataset reconstructs 1-km warm- and cold-season distributions of cattle and sheep/goats across China from

Read the Wild Lands assessment →

Keep exploring

Follow the systems. Find the connections.

WildlifeLandWaterOceansClimateCitiesOne HealthGovernanceMethods / Monitoring
Browse the archive →