GovBot Tides & Terrain

View saved version →

---
name: terrain-watershed-brief
description: Prepare a public-point terrain and watershed context brief using USGS elevation and Watershed Boundary Dataset records. Use for modeled ground elevation, nested HUC context and source limitations, not surveys, elevation certificates, parcel flood zones or station-datum conversions.
---

# Terrain and watershed brief

Combine two kinds of geographic evidence while keeping modeled elevation and watershed membership separate from parcel-level determinations.

1. Use a user-supplied public research point or a clearly identified public point already in the request. Do not ask for a home address, obtain device location, invent coordinates or offer address lookup. Confirm latitude/longitude and units; reject invalid coordinates before a call rather than treating an error as no data.
2. Call `get_ground_elevation_at_point` with the point and explicit units. Read the returned elevation, status/no-data indicator, vertical reference, source dataset and resolution/vintage fields where supplied. Preserve source labels and missing metadata. A sampled elevation model value is not a field survey, building-floor elevation or certified site elevation. Resolution is not vertical accuracy, and retrieval date is not terrain survey date.
3. Call `get_watershed_at_point` for the requested levels (available HUC2/4/6/8/10/12; default 8/10/12). Keep HUC identifiers as strings with leading zeros, retain names and level labels, and distinguish nested units from separate unique watersheds. Retain all multiple boundary matches and any hierarchy inconsistency. Report the returned spatial relation and source coverage; do not infer flow direction, a drainage path, hydraulic connectivity, floodplain or flood risk from membership alone.
4. Present the point and its stated public context; modeled elevation with units and only documented vertical reference; a compact HUC table; source links and separate source/retrieval/cache dates; and any no-data, unmatched level or missing source metadata. Do not convert a sentinel, null, unsupported point, source failure or absent polygon into elevation zero, no watershed or safety.
5. When asked to combine with tides, explain that a coastal station's tidal-to-geodetic datum relationship is local to that station and epoch. Do not add or subtract its offset from arbitrary inland EPQS terrain or claim the two references are interchangeable. A defensible local conversion requires appropriate authoritative transformation/control evidence beyond these tools.

Decline property flood-zone, engineering, insurance, safety or elevation-certificate conclusions. Offer this source-qualified regional context and direct requests for a certified measurement or determination to the relevant qualified professional/authority; do not promise that a missing record excludes a hazard.

View saved version →

---
name: tide-prediction-brief
description: Prepare a NOAA astronomical tide prediction brief for a station and bounded dates, or compare its predictions with matching observations. Use when users need tide times, predicted levels and the meaning of the selected datum, units and local clock, with station capability and coverage limits.
---

# Tide prediction brief

Make a usable, dated tide table or a qualified observation comparison without turning astronomical predictions into a weather or flooding forecast.

1. Confirm station or public research point, dates, desired high/low table versus sampled series, units, datum and time zone. Use `find_tide_stations_near_point` with `station_type: tide_predictions` when station selection is needed. Explain distance and capability; do not substitute a nearby gauge silently or solicit a home address/device location.
2. Check `get_station_datums` and station capability metadata. Preserve station datum labels, units and published epoch; absent datum/epoch stays unknown. Known subordinate stations require high/low predictions and MLLW in this service. Do not force another interval or datum from an unsupported station.
3. Use `get_tide_predictions` with explicit arguments: `high_low` allows at most 31 calendar days, `hourly` seven, and `six_minute` two. State any narrowed window or product change. `local` follows station local/daylight-saving convention, `local_standard` does not adjust for daylight saving, and `gmt` is UTC. Keep the returned clock label next to times; do not invent missing UTC offsets or silently convert ambiguous DST times.
4. Present requested versus returned coverage, high/low type or sampling interval, predicted height and exact datum/units, station/source link and retrieval/cache provenance. Distinguish source dates and prediction period from fetch time. Null/failed/missing data are not zero tides. Do not label astronomical predictions as observations, water-level weather forecasts, storm-surge predictions or safety advice.
5. For a requested comparison, call `get_observed_water_levels` using the same station, datum, units, clock and compatible interval. Pair only matching timestamps with numeric values; give the number of pairs and unmatched/missing rows. Do not pair repeated local DST clock times ambiguously; use a compatible unambiguous time basis or leave the comparison unresolved. Explain observation quality flags and preliminary versus verified status. Label observed-minus-predicted differences with units and scope; do not call them an independently established storm surge, causal weather effect or site flood depth.

Conclude with what the table supports: station-level astronomical tide research on the requested dates. No transfer of station datum offsets to arbitrary inland ground, no property certification, flood-zone or insurance determination, and no guarantee of safe boating or flood-free conditions. For operational decisions, point to the responsible official service without inventing a live alert check.

View saved version →

---
name: water-level-history-brief
description: Prepare a dated NOAA coastal water-level observation brief for a named station or public research point, including station suitability, reported levels, quality flags, datum and time coverage. Use for historical measured water levels, not astronomical tide predictions or property flood determinations.
---

# Water-level history brief

Turn a bounded observation request into a source-linked account of what a coastal gauge measured and what the record can support.

1. Establish the station or public research point, dates, units, vertical datum and clock convention. Do not solicit a home address, device location or personal coordinates; there is no address lookup. When location is supplied, use `find_tide_stations_near_point` with `station_type: water_level_observations`. Report distance and suitability; the nearest gauge is not a measurement at the research point. Preserve the discovery radius, displayed count and limits. Discovery covers current active water-level and tide-prediction catalogs, not every historical station.
2. Use `get_station_datums` for the chosen station to explain the requested datum, station metadata and published epoch. A datum absent from the station record is not available merely because its name is in the tool schema. Preserve returned units and labels: datum heights and tidal ranges are lengths, but HWI/LWI are hour intervals. Do not transfer station offsets to terrain or other locations.
3. Call `get_observed_water_levels` with explicit dates, datum, units and time zone. `six_minute` accepts at most four calendar days; `hourly_verified` and `high_low_verified` accept at most 31. For longer questions, agree a bounded interval or describe a finite set of requested windows before querying; do not imply a short sample covers a whole month. Never substitute a verified product for six-minute data without stating the change.
4. Inspect actual returned coverage, missing values, flags and preliminary/verified status. Product names do not justify marking every row verified. Keep unavailable evidence, no records and an observed zero distinct. Preserve time-zone labels: `gmt` is UTC; `local_standard` excludes daylight-saving time, while `local` follows the station's local daylight-saving convention. Do not invent UTC offsets at DST transitions.
5. Report a brief with station identity/source link; requested versus returned dates and sampling; datum/epoch/units/clock; selected levels or extrema with timestamps and quality; coverage and interpretation limits. Calculate extrema only from numeric returned rows and label them as sample extrema, not a complete-period record. Distinguish source data timestamps, actual retrieval and cache age/expiry when returned. A cache hit is not a fresh source observation.

If comparing predictions, use the tide-prediction workflow and match station, interval, datum, units and clock before pairing rows. Observed water level includes influences beyond astronomical tides; a difference is not by itself a validated surge estimate or causal attribution. Do not certify flood zones, building safety, navigation safety, insurance eligibility or property elevation. Provide the public evidence and explain the missing basis for such a determination.