diff --git a/docs/measuring-point-height-null-handling.md b/docs/measuring-point-height-null-handling.md
new file mode 100644
index 00000000..0611e580
--- /dev/null
+++ b/docs/measuring-point-height-null-handling.md
@@ -0,0 +1,26 @@
+# Measuring-point height: null handling
+
+**Status:** Accepted
+**Date:** 2026-08-26
+**Decided by:** Cris Morton + Ethan Mamer
+**Recorded by:** Kelsey Smuczynski
+
+## Context
+
+Three OGC layers compute depth-to-water and water-elevation values from measuring-point height:
+
+- `water_well_summary`
+- `depth_to_water_trend_wells`
+- `water_elevation_wells`
+
+Each uses `COALESCE(measuring_point_height, 0)` in the calculation. When `measuring_point_height` is null, this treats the measuring point as ground surface level (0), rather than excluding the record or flagging it for review.
+
+## Decision
+
+A missing or null measuring-point height defaults to zero (ground surface level) in these calculations. Cris Morton confirmed this is the correct default on 2026-08-26.
+
+## Consequences
+
+- Depth-to-water and water-elevation values for wells with no recorded measuring-point height are calculated as if the measurement were taken at ground surface. These values carry more uncertainty than wells with a known measuring point, but nothing in the output currently distinguishes them.
+- If a future review determines this default introduces meaningful error for a subset of wells (for example, wells with a well casing or stickup that's unusually tall), revisit this decision and consider flagging affected records instead of silently defaulting to zero.
+- Any new analytic layer that calculates depth-to-water or water-elevation from measuring-point height should follow this same convention, or explicitly document why it doesn't.