[DOCS] Add bathymetry guide page documenting deeper-than coverage semantics

docs/guides/base/bathymetry.mdx

@@ -0,0 +1,42 @@
+---
+title: Bathymetry
+description: Underwater depth features in the base theme and how their coverage is defined
+pagination_label: Base Bathymetry
+---
+
+# Bathymetry
+
+The base theme's **bathymetry** feature type provides topographic representations of underwater areas, such as parts of lake beds or ocean floors. The geometries are derived from vectorized bathymetric data products built from [ETOPO1](https://www.ncei.noaa.gov/products/etopo-global-relief-model) and [GLOBathy](https://www.nature.com/articles/s41597-022-01132-9) data.
+
+Bathymetry features carry Polygon or MultiPolygon geometries and a single classifying property, `depth`, which gives the depth level of the feature in meters below the water surface. See the [bathymetry schema reference](/schema/reference/base/bathymetry/) for the full column definitions.
+
+## Stacked depth levels
+
+Bathymetry is delivered as a set of **stacked depth levels**. Each depth level is a coverage polygon set: the features with `depth = d` together cover every point of the earth's surface where the water is **deeper than** `d` meters.
+
+Practically, this means entities with a lower `depth` will always totally cover entities with a higher `depth`:
+
+- `depth = 0` covers all water — oceans and the lakes captured by GLOBathy.
+- `depth = 200` covers only the water that is deeper than 200 m.
+- `depth = 4000` covers only the water that is deeper than 4000 m.
+
+A point on the ocean floor at 3,500 m depth is therefore covered by every level from 0 up to the deepest level shallower than 3,500 m. Land is not covered by any level.
+
+To render a classic bathymetric tint map, draw the levels stacked from shallow to deep: paint `depth = 0` first in the lightest color, then paint each successively deeper level on top in a darker color. The `cartography.sort_key` property encodes this recommended stacking — features with a lower `sort_key` should be drawn in front of features with a higher `sort_key`. To extract a single disjoint depth *band* instead (for example, water between 200 m and 1000 m deep), subtract the deeper level's coverage from the shallower level's coverage.
+
+## Takeaways
+
+- **Land is not covered.** Bathymetry polygons only exist over water; no masking against land or water layers is needed.
+- **Geometries are grid chips.** Instead of a small number of world-spanning multipolygons, each depth level is delivered as many polygons chipped to a roughly 1° grid, the same way Overture's ocean water polygons are delivered. Spatial filters (for example, `bbox` predicates) are much more effective as a result.
+- **Semantics are "deeper than."** A feature with `depth = d` means "the water here is deeper than `d` meters," and levels overlap by design. Count on multiple features covering the same point — one per depth level shallower than the water at that point.
+
+## Data access
+
+Bathymetry data is available at:
+
+| provider | location |
+| --- | --- |
+| Amazon S3 | `s3://overturemaps-us-west-2/release/<release>/theme=base/type=bathymetry/*` |
+| Azure Blob Storage | `https://overturemapswestus2.blob.core.windows.net/release/<release>/theme=base/type=bathymetry/*` |
+
+See the [base guide overview](./index.mdx) for general information about the base theme, including licensing and release cadence.

docs/guides/base/index.mdx

@@ -20,8 +20,9 @@ Some of our schema design choices include:
 
 ## In this guide
 
-This overview covers the feature types, dataset construction, and access for the base theme. A dedicated page covers one feature type in more depth:
+This overview covers the feature types, dataset construction, and access for the base theme. Dedicated pages cover some feature types in more depth:
 
+- **[Bathymetry](./bathymetry):** Underwater depth features — how the stacked "deeper-than" depth levels work, how the coverage is built, and what it means for rendering and querying.
 - **[Infrastructure](./infrastructure):** Transit features such as bus stops, platforms, and stations — including how they relate to the places and transportation themes — and entities like mountains that appear in both base and places by design, with example queries.
 
 ## Dataset description
@@ -31,7 +32,7 @@ This overview covers the feature types, dataset construction, and access for the
 There are six feature types in the base theme:
 
 #### Bathymetry 
-Bathymetry features provide topographic representations of underwater areas, such as parts of lake beds or ocean floors. The geometries are derived vectorized bathymetric data products from [ETOPO1](https://www.ncei.noaa.gov/products/etopo-global-relief-model) and [GLOBathy](https://www.nature.com/articles/s41597-022-01132-9) data.
+Bathymetry features provide topographic representations of underwater areas, such as parts of lake beds or ocean floors. The geometries are derived vectorized bathymetric data products from [ETOPO1](https://www.ncei.noaa.gov/products/etopo-global-relief-model) and [GLOBathy](https://www.nature.com/articles/s41597-022-01132-9) data. Bathymetry is delivered as stacked depth levels, where each feature with `depth = d` covers all water deeper than `d` meters; see the [bathymetry page](./bathymetry) for details.
 
 #### Infrastructure 
 Infrastructure features provide basic information about real-world infrastructure entities such as bridges, airports, runways, aerialways, communication towers, and power lines. All of the features are pulled from OpenStreetMap and converted to the Overture schema. See the [infrastructure page](./infrastructure) for a closer look at transit features and how the infrastructure type relates to other themes.

sidebars.js

@@ -58,6 +58,7 @@ const sidebars = {
           collapsed: true,
           items: [
             'guides/base/index',
+            'guides/base/bathymetry',
             'guides/base/infrastructure',
           ],
         },