Landslide susceptibility (global, NASA, slope-gated 90 m)
The NASA landslide susceptibility classes carried onto a 90 m grid and zeroed where the terrain is too flat for a slide to start.
Source: NASA Global Landslide Susceptibility (Stanley & Kirschbaum) × GIRI / NGI maximum terrain slope Year: 2017 Category: Physical risks · Mitigating services · Geological Coverage: Global Format: Raster grid (92.8 m) Used in risk analysis: No (contextual layer; not used in materiality scoring) — the flag and the priced loss read Landslide susceptibility (global, NASA)
What it shows
The same six-level NASA landslide susceptibility scale as the global NASA layer (0 None, 1 Very low, 2 Low, 3 Moderate, 4 High, 5 Very high), carried onto a 90 m grid and set to 0 wherever the terrain is too flat for a slide to begin.
The point of the layer is the slope gate. A susceptibility class built from terrain, geology and land cover spreads over a whole neighbourhood; the ground that can actually fail is the steep part of it. At 2 km, a valley floor and the slope above it share one pixel and one class. At 90 m with the gate applied, the flat floor reads 0 and the slope keeps its class.
How it is built
The NASA susceptibility raster is carried onto a 92.766 m web-mercator grid (EPSG:3857, Float64), and every cell whose maximum terrain slope is below 6 degrees is set to class 0. The slope comes from the GIRI maximum terrain slope surface (UNEP-GRID / CIMA / NGI).
Two consequences follow from that construction, and both matter when reading a value.
Class 0 means "none or too flat". It merges two different statements: NASA assigning no susceptibility, and the slope gate zeroing a cell NASA did classify. The legend label says so, and the 2 km layer's class 0, which reads simply "None", is not the same quantity.
It is initiation only. The gate answers where a slide can start. Runout is not modelled, so a flat cell immediately below a steep, high-class slope reads 0 while remaining exposed to material arriving from above. That is the main way this layer can be read too favourably.
How to read it
Higher classes mean a greater inherent likelihood that the ground fails at that location. The classes are an ordinal hazard scale from NASA, not a probability and not a physical quantity in units.
This is a screening layer, not a siting tool. The slope gate sharpens the picture rather than certifying it: below 6 degrees it removes the class, and above 6 degrees it leaves NASA's own classification untouched, at a resolution the source classification was never built at. Reading a single 90 m cell as a statement about a specific parcel asks the underlying map for more than it carries.
How it differs from the 2 km layer
Both layers carry the same NASA classification, and the catalogue keeps both.
| Landslide susceptibility (global, NASA) | This layer | |
|---|---|---|
| Grid | ~2 km (mode resampling of the native ~1 km raster) | 92.8 m |
| Slope gate | None | Class set to 0 below 6 degrees of maximum terrain slope |
| Class 0 reads | "None" | "None or too flat" |
| Proximity flag | Yes, Mitigating services at class 3 and above | None |
| Read by the loss engine | Yes | No |
The 2 km layer is the operating one: it raises the "Soil and sediment retention" flag at Moderate and above, and it is the raster the Direct Physical Loss engine samples for ground movement. This 90 m layer is published for reading, so that a flag raised at 2 km can be inspected at a resolution where slope is visible. Nothing is computed from it today, and moving the flag onto it would change which sites are flagged, which is a decision for the layer team rather than a by-product of the finer grid.
Class thresholds
This is a categorical susceptibility layer. It publishes six ordinal classes and no proximity rule, so it crosses no threshold in the platform.
| Value | Class |
|---|---|
| 0 | None or too flat |
| 1 | Very low |
| 2 | Low |
| 3 | Moderate |
| 4 | High |
| 5 | Very high |
How these thresholds were set. The six-class scheme and the rules that assign a cell to a class belong to NASA's Global Landslide Susceptibility Map methodology (Stanley & Kirschbaum) and are not reproduced in Darwin's configuration. The in-house elements are the 90 m grid and the 6 degree slope gate, a Darwin cut-off on the GIRI maximum-slope surface rather than a published landslide-initiation standard.
Reference: NASA Global Landslide Susceptibility Map (Stanley & Kirschbaum; Stanley et al., 2017).
Source
NASA Global Landslide Susceptibility Map (Stanley & Kirschbaum, 2017). Maximum terrain slope from GIRI (UNEP-GRID / CIMA / NGI), CC BY 3.0 IGO.
Comparison with the WWF Risk Filter Suite
This layer maps to WWF Biodiversity Risk Filter indicator S3_1 Landslides and draws on the same NASA source family (Stanley et al., 2017). It departs from WWF in two ways: the classes are carried at 90 m rather than at the coarse grid WWF publishes, and the slope gate removes ground the WWF indicator would still score.
Legend
Symbolised field: Landslide susceptibility
| Value (Landslide susceptibility) | Label | Colour |
|---|---|---|
| 0 | None or too flat | #ffffcc |
| 1 | Very low | #d9f0a3 |
| 2 | Low | #addd8e |
| 3 | Moderate | #fdae61 |
| 4 | High | #f46d43 |
| 5 | Very high | #d73027 |
Generated from darwin/layers/layer-landslide-slope-resolved.toml (develop).