Lower Uncompahgre River

Division 4 · Water District 41

Draft - NLDI/CDSS verified

Bootstrap review report — d4_wd41_lower_uncompahgre_river

Generated 2026-08-13 22:45 UTC for Division 4 / Water District 41.

This is a review checklist, not a finished config. Everything below is a suggestion from live DWR data — confirm or correct each section in config.yaml before running this basin for real.

1. Stream-name patterns

Assumed mainstem pattern: UNCOMPAHGRE RIVER (highest-count waterSource value). Candidates below by structure count — decide which are real tributaries worth their own centerline vs. noise, and add them to longitudinal.tributary_stream_patterns in priority order (more-specific names, e.g. "WEST FORK X", must precede substrings of themselves, e.g. "X").

waterSource	structure count
UNCOMPAHGRE RIVER (assumed mainstem)	28
SPRING CREEK	10
HAPPY CANYON CREEK	7
DRY CREEK	6
HORSEFLY CREEK	5
CEDAR CREEK	4
BEATON CREEK	3
GUNNISON RIVER	3

2. Diversion / return classification

Auto-classified from structureType + name patterns: 71 diversions, 0 returns, 5 unclassified (neither pattern matched — these render as neither on the schematic unless added to return_flow_wdids/force_diversion_wdids or a type is added to the diversion/return type lists).

This layer needs the least manual work — it's the same heuristic the engine uses at runtime. Spot-check a sample rather than reviewing every row; if you have a water-year accounting spreadsheet for this basin, cross-reference it against the 'other' bucket specifically (that's where a real augmentation return with an unusual DWR structureType will hide).

2b. Cross-checked against decreed water rights (CDSS)

None flagged — no unclassified structure's decreed water right clearly disagreed with its current classification (checked up to the API-budget cap; see bootstrap.py's cross_check_classification for the cap).

3. Suspected nested tributaries

Still need a human decision — these tributary pairs are comparably close to EACH OTHER as to the mainstem — one may actually join the other rather than the mainstem directly (like Leavenworth Creek -> South Clear Creek in the Clear Creek basin). NLDI couldn't resolve these (no nearby flowline, or ambiguous/no downstream overlap within range) — if you know the real topology, set it in longitudinal.tributary_parent_overrides.

tributary	possible parent	dist to possible parent (mi)	dist to mainstem (mi)	NLDI result
BEATON CREEK	GUNNISON RIVER	3.93	2.92	NLDI check failed: 429 Client Error: Too Many Requests for url: https://api.water.usgs.gov/nldi/linked-data/comid/position?f=json&coords=POINT%28-107.731513+38.351644%29
GUNNISON RIVER	CEDAR CREEK	0.09	6.8	NLDI check failed: 429 Client Error: Too Many Requests for url: https://api.water.usgs.gov/nldi/linked-data/comid/position?f=json&coords=POINT%28-107.743605+38.484104%29

4. Gage placement risk

5 gage(s) would currently fail placement (> 2.5mi from the nearest centerline anchor — same failure mode as the Clear Creek at Derby bug). Options: raise the cutoff (check it doesn't sweep in an unrelated stream first — see the Little Dry Creek case), or add a manual override.

gage	name	nearest stream	distance (mi)
UVTFVLCO	IRONSTONE CANAL TAIL NEAR PEA GREEN, CO	DRY CREEK	4.5
UVTCQLCO	CQ LATERAL TAIL NEAR PEA GREEN, CO	DRY CREEK	4.39
UVTFQLCO	FQ LATERAL TAIL NEAR PEA GREEN, CO	DRY CREEK	3.38
UVTECNCO	EAST CANAL TAIL AT DRAGON'S TEETH DROP, CO	MAINSTEM	3.03
UVTMDCCO	MONTROSE & DELTA CANAL TAIL ABOVE STATE HWY 348, CO	DRY CREEK	2.5

30 more gage(s) pass but are within 2x the cutoff distance — worth a glance if the structure dataset changes later.

Next steps

Fill in tributary_stream_patterns from section 1.
Resolve section 3's flags (or confirm none apply).
Resolve section 4's at-risk gages.
Cross-reference return_flow_wdids/force_diversion_wdids against any local accounting source you have for this basin.
Review the rendered schematic for this district before relying on it.

Reconstruct conditions as of

Evolution of the network

Computed 2026-08-20 08:57 from 30 gage(s) with cached period-of-record data. Statistical tests (Mann-Kendall trend, Pettitt changepoint) require at least 8 years spanning at least 10 years of record -- gages below that report no signal, not "no change."

Data provenance & quality

Where these numbers come from and how far they can be trusted. Grade C (3/6 points) — the components below are the finding; the letter alone is not. Sources: DWR HydroBase snapshot HydroBase_CO_20260707 (imported 2026-08-19) for history, plus the CDSS REST API for current conditions.

Check	State	What was measured
gage freshness	stale	newest gage reading is 45 day(s) old (target 7)
diversion record	expected-lag	newest diversion record is 293 day(s) old (DWR publishes ~1 year in arrears)
structure coverage	sparse	7% of schematic structures have data

Independent check: PASS — 100.0000% agreement across 291,280 stored values re-checked against DWR's authoritative record (HydroBase_CO_20260707), over 20 structures. 0 disagreement(s) beyond a 0.01 cfs tolerance. Checked 2026-08-19; a rotating sample covers the whole basin across successive runs, and the structures this report cites are checked every time.

Federal cross-check: PASS — 97.48% day-level agreement with USGS's own published records for the same gages, across 2 co-listed station(s) and 36,337 overlapping days (within max(2%, 0.5 cfs) daily). Pre-1950 records at some gages carry known between-agency day-stamp and transcription differences — scored at each year's best ±1-day alignment; residual disagreement is real historical divergence, flagged for review.

30 gage(s) / 148,598 daily values (1938-10-01–2026-07-06); 133 structure(s) with diversion records; 1104 with decreed rights. Methodology has not yet had independent expert review — treat every figure as an advisory screen, not an administrative or legal conclusion.

Read this first: the measurement network itself evolved

This basin was not always measured the way it is today — the API-era record is dense and recent, the early record sparse. 28 gage(s) and 61 structure(s) report in 2026, but structure reporting only reached half its modern level around ~1950. Every historical claim below is checked against this: long-term shifts that co-move with the growth of the network carry an explicit measurement-evolution caution, and abandonment screens whose idleness begins before the modern reporting era are flagged †.

Decade	Gages with data (median yr)	Structures reporting (median yr)
1900s	0	1
1930s	1	2
1940s	1	0
1950s	1	57
1960s	1	70
1970s	1	74
1980s	1	72
1990s	4	68
2000s	5	52
2010s	15	63
2020s	23	61

Gage flow trend & changepoints

Every located gage with a period of record — dot size scales with the strength of its flow trend (|Sen slope| in %/decade, the same statistic as the table below), color with direction. An amber ring marks a detected Pettitt changepoint. Click a gage to chart its trend and shift.

Gage	Trend	Change / decade	Significance	Years	Changepoint
▸ SELIG CANAL SELGCNCO	↓ decreasing	-148.0%	p=0.018	15	~2019
▸ UNCOMPAHGRE RIVER UPSTREAM OF SOUTH CANAL UNCUPSCO	↓ decreasing	-48.3%	p=0.036	17	~2017
▸ ABC LATERAL ABCLATCO	↓ decreasing	-39.6%	p=0.000	40	~2013
▸ SOUTH CANAL NEAR MONTROSE SOUCANCO	↑ increasing	+23.5%	p=0.002	36	—
▸ MONTROSE & DELTA CANAL MANDCNCO	↓ decreasing	-13.8%	p=0.003	15	—
▸ UNCOMPAHGRE RIVER AT DELTA, CO. UNCDELCO	↑ increasing	+2.8%	p=0.048	88	~1981

24 other gage(s) show no significant trend or changepoint (not shown).

Mass-balance (network closure) trend

Structure divrec-history backfill coverage as of the last check: 7% of 460 schematic-relevant structures. 0 year(s) clear the 30% coverage gate so far -- the trend activates automatically once qualifying years span a full decade. No action needed; the nightly backfill keeps extending the record.

Reach gain/loss shifts (1991–2026)

Mass-balance kinks computed for every summer month back to 1991 (854 reach-snapshots) — which reaches gain or lose water, and how that has shifted across decades. Each shift is cross-checked against the observation record above: as more diversions became metered, "unaccounted" water shrinks mechanically, which can masquerade as a hydrologic shift.

FAIRVIEW INLET OFF THE SOUTH CANAL NEAR MONTROSE: reach shifted from -350 cfs to -95 cfs summer mean (more gaining, p=0.000, 17 yr) [alluvial valley: real alluvial aquifer -- a decades-long shift toward losing is the classic signature of alluvial pumping and urbanization] [1 alluvial wells within 2 mi of the gage; pooled water table -0.5 ft (2014–2024, decreasing)]
UNCOMPAHGRE RIVER NEAR OLATHE: reach shifted from +108 cfs to -187 cfs summer mean (more losing, p=0.002, 25 yr) [alluvial valley: real alluvial aquifer -- a decades-long shift toward losing is the classic signature of alluvial pumping and urbanization]
GARNET DITCH: reach shifted from -63 cfs to -18 cfs summer mean (more gaining, p=0.000, 16 yr) [alluvial valley: real alluvial aquifer -- a decades-long shift toward losing is the classic signature of alluvial pumping and urbanization] [measurement-evolution caution: this shift co-moves with the growth of the gage/metering network (rho=-0.69) -- part of it may be unaccounted water becoming accounted, not hydrologic change] [3 alluvial wells within 2 mi of the gage; pooled water table -0.3 ft (2014–2024, no_trend)]
IRONSTONE CANAL NEAR OLATHE, CO: reach shifted from +124 cfs to +164 cfs summer mean (more gaining, p=0.010, 16 yr) [alluvial valley: real alluvial aquifer -- a decades-long shift toward losing is the classic signature of alluvial pumping and urbanization] [3 alluvial wells within 2 mi of the gage; pooled water table +1.2 ft (1976–2024, no_trend)]

Groundwater now has its own tab: 76 monitored well(s), water-table trends, stream corridors, and the advisory depletion screen.

Reservoirs now has its own tab: 3 reservoir(s), end-of-year storage back to 1950, click-to-chart histories.

Cross-gage pattern fingerprint

Over the 35-year window with full coverage (1992–2026), the dominant shared pattern across 3 gage(s) explains 63% of year-to-year variance (almost certainly a basin-wide climate/snowpack driver).

ABCLATCO doesn't move with that shared pattern -- a real, localized signal (a new diversion, land-use change, or a gage-specific data problem) worth a closer look, not necessarily noise.

The basin's aquifer story, from 8,080 water-level measurements across 76 monitored well(s) (64 alluvial, 0 bedrock, rest unclassified) — plus 0 administered well(s) reporting metered pumping through diversion records. Observation wells and pumping wells are different populations and are never conflated. Click a well to chart its record.

Well map

Every monitored well with a location — hover for identity, click to chart its water-level record below. ● alluvial ● bedrock / confined (Denver Basin, High Plains, etc. — one-way, no human-timescale recharge) ● unclassified — marker size scales with the strength of the well's water-level trend (|Sen slope| in ft/decade — the same Mann-Kendall/Sen statistic as the trends table below); the smallest, faintest dots have too little record to test.

3D aquifer x-ray

Every aquifer unit as its own 3D head surface — alluvial water table on top, Denver Basin members (Dawson, Denver, Arapahoe, Laramie–Fox Hills) stacked beneath where measured — interpolated from measured water levels and animated through time. Units are never mixed; surfaces exist only inside each unit's well coverage (no extrapolation); thin epochs are suppressed with their reason shown. Vertical scale exaggerated for readability; z is head elevation in feet. Advisory — measured heads, not a calibrated model.

Water-table trends

Of 28 well(s) with enough record for a trend test: 6 declining, 2 rising.

Fastest-declining wells	Setting	ft / decade	Span	p
▸ LGRB-Poly 30	alluvial	-2.87	10 yr	0.0491
▸ NC04900936BA LGRB Poly 28	alluvial	-2.79	11 yr	0.0051
▸ LGRB-Poly 23	alluvial	-2.74	10 yr	0.0042
▸ NC05101035BD LGRB Poly 8	alluvial	-2.64	12 yr	0.0236
▸ LGRB-Poly 19	alluvial	-2.1	10 yr	0.02
▸ LGRB-Poly 6	alluvial	-1.44	10 yr	0.0491

Stream corridors — the surface connection

For each reach whose gain/loss has shifted (Trends tab), the wells within 2 miles of its anchoring gage, their pooled water-table history over the same era, and an advisory Glover screen of lagged stream depletion from nearby metered pumping. v1 proximity = distance to the reach's anchoring gage; segment-based reach join is a flagged upgrade.

FAIRVIEW INLET OFF THE SOUTH CANAL NEAR MONTROSE: 1 alluvial well(s) nearby; pooled water table -0.5 ft (2014–2024, decreasing, p=0.0491)

UNCOMPAHGRE RIVER NEAR OLATHE: 3 alluvial well(s) nearby

GARNET DITCH: 3 alluvial well(s) nearby; pooled water table -0.3 ft (2014–2024, no_trend, p=0.5915)

IRONSTONE CANAL NEAR OLATHE, CO: 3 alluvial well(s) nearby; pooled water table +1.2 ft (1976–2024, no_trend, p=0.0584)

Wells labeled by administrative world before anything else is said about them: tributary alluvium participates in the priority system; designated-basin and Denver Basin bedrock aquifers are governed differently and are excluded from stream-connection reasoning. Read bedrock declines differently too: deep bedrock aquifers recharge negligibly on human timescales, so a falling bedrock level is storage mining — effectively permanent — not a drought response that a wet year undoes. All screens advisory.

3 reservoir(s) with 58 storage measurement(s), 1950–2022. Storage change is real water the mass balance must credit — a losing reach beside a filling reservoir isn't losing water, it's banking it.

Reservoir map — how full is each vessel?

Each reservoir drawn as a vessel: bubble size ∝ its record-high storage, fill level = last recorded storage as a share of its surveyed capacity (USACE National Inventory of Dams) where a confident NID match exists — the tooltip names the matched dam — and of its own record high otherwise (a wrong capacity would be worse than an honest record high, so only confident joins upgrade). Hover for numbers, click to chart its history below the map. A dashed teal ring marks reservoirs reporting live DWR telemetry — their fill reflects yesterday's storage, drawn as a dotted tail on charts and treated as provisional until the next verified HydroBase snapshot supersedes it. ● <25% ● 25–60% ● 60–90% ● >90%

Basin storage, end of each water year

End-of-water-year totals; per-year reservoir counts shown because reporting coverage varies -- a total is only comparable to years with similar counts.

Largest reservoirs on record

Click a reservoir to chart its full storage history.

Reservoir	Last recorded storage (AF)	As of
▸ CITIZENS RESERVOIR (4103651)	123	1956-09-01
▸ MOCK RESERVOIR NO 1 (4103653)	40	1952-05-01
▸ BUCKHORN RESERVOIR #2 (4103650)	0	2022-09-02

Decrees, seniority, and administrative calls connected to this basin's flow network. Advisory reading aids on public DWR data — not administrative or legal conclusions. Rights on file: 1249 across 1104 structure(s); 72 structure(s) carry 0 court-case references.

Abandonment radar (25 screens)

Absolute rights with 10+ years of no recorded diversions. Screens only — storage, alternate points, and plan operations can explain nonuse lawfully. And diversion REPORTING itself only reached its modern level ~1950 — idleness that begins before then may be unreported use, not nonuse (flagged † below).

E E FRASIER SPGS & SEPG (4100616) (1887-06-01, 1.5 cfs decreed): last active 1950 — 75 years idle
SPRING CREEK NO1 PL (4100711) (1976-09-07, 0.1 cfs decreed): last active 1954 — 71 years idle
SWEITZER DITCH (4100714) (1911-09-02, 1.0 cfs decreed): last active 1954 — 71 years idle
MONTROSE CITY DITCH (4100683) (1884-04-05, 10.0 cfs decreed): last active 1955 — 70 years idle
EAST FORK FDR PIPELINE (4100625) (1933-10-25, 1.0 cfs decreed): last active 1956 — 69 years idle
LABARR DITCH NO1 (4100664) (1976-04-29, 1.8 cfs decreed): last active 1969 — 56 years idle
CHAPARRAL DITCH (4100603) (1911-09-02, 0.4 cfs decreed): last active 1974 — 51 years idle
MENKE NORTH DITCH ENLT (4100686) (1947-03-01, 1.0 cfs decreed): last active 1974 — 51 years idle
CITIZENS RESERVOIR (4103651) (1912-09-01, 0.0 cfs decreed): last active 1974 — 51 years idle
MOCK RESERVOIR NO 1 (4103653) (1911-09-02, 0.0 cfs decreed): last active 1974 — 51 years idle

Compliance screens (84 advisory flags)

Screened 112 of 1104 structures with rights (118 have diversion history on file). A flag is a lead for expert review — augmentation plans, exchanges, and free-river conditions can make over-decree diversions lawful and are deliberately not modeled.

IRONSTONE CANAL (4100534): Observed p99 daily rate 456.0 cfs (max 578.0) exceeds total decreed absolute 202.7 cfs across 48877 recorded day(s).
SELIG CANAL (4100559): Observed p99 daily rate 260.0 cfs (max 377.0) exceeds total decreed absolute 121.9 cfs across 51459 recorded day(s).
EAST CANAL (4100520): Observed p99 daily rate 189.0 cfs (max 1530.0) exceeds total decreed absolute 60.4 cfs across 45736 recorded day(s).
LOUTSENHIZER CANAL (4100537): Observed p99 daily rate 188.0 cfs (max 741.0) exceeds total decreed absolute 46.4 cfs across 46945 recorded day(s).
IRONSTONE EXTENSION D (4100983): Observed p99 daily rate 166.0 cfs (max 192.0) exceeds total decreed absolute 61.2 cfs across 2128 recorded day(s).
UNCOMPAHGRE CEDAR CR D (4100571): Observed p99 daily rate 44.1 cfs (max 58.0) exceeds total decreed absolute 29.1 cfs across 9853 recorded day(s).
F. J. LATERAL (4100631): Observed p99 daily rate 36.8 cfs (max 39.5) exceeds total decreed absolute 15.0 cfs across 2078 recorded day(s).
ROSS BROS DITCH (4100554): Observed p99 daily rate 34.9 cfs (max 38.3) exceeds total decreed absolute 22.2 cfs across 26028 recorded day(s).
STARK VOLKMAN DITCH (4100566): Observed p99 daily rate 30.3 cfs (max 55.3) exceeds total decreed absolute 20.1 cfs across 32500 recorded day(s). Structure holds an Augmentation-use decree -- diversions above the direct-flow decree can be lawful under plan operations; interpret with care.
CHIPETA BEAUDRY DITCH (4100515): Observed p99 daily rate 23.6 cfs (max 32.0) exceeds total decreed absolute 20.8 cfs across 28093 recorded day(s).

Seniority (top structures by decreed rate)

Structure	Senior priority date	Rights	Decreed abs (cfs)	Conditional?
MONTROSE & DELTA CANAL (4100545)	1881-11-01	11	627.1	—
IRONSTONE CANAL (4100534)	1882-01-05	7	202.7	—
DRY CREEK FEEDER D (4100610)	1887-02-28	1	180.0	—
SELIG CANAL (4100559)	1882-01-27	6	121.9	—
GARNET DITCH (4100527)	1883-06-18	2	93.3	—
UNCOMPAHGRE (LOUTSENHIZE (4101233)	1882-02-23	3	81.6	—
SPRING CREEK VALLEY D (4100564)	1883-06-13	1	65.1	—
IRONSTONE EXTENSION D (4100983)	1885-02-28	1	61.2	—
EAST CANAL (4100520)	1882-05-01	4	60.4	—
LOUTSENHIZER CANAL (4100537)	1881-11-01	4	46.4	—
PURDY AND VICKERS DITCH (4100517)	1882-03-02	5	46.2	—
HAIRPIN LTL D (4100532)	1911-09-02	1	45.0	—
SATISFACTION DITCH (4100558)	1882-01-27	3	36.0	—
OURAY DITCH (4100549)	1877-05-01	10	36.0	—
BALDY DITCH (4100501)	1884-06-20	1	30.0	—

No inventory or legal-document changes detected since the daily watch began.

If a call were placed at a given seniority, which structures in this basin would be out of priority, and how much of their decreed rate would be curtailed? Computed from this basin's cached decrees.

What this does and doesn't model. Seniority arithmetic only: every right junior to the call is treated as curtailed. It does not model reach applicability — structures downstream of the calling point are not administered by it — and it applies no futile-call judgment. Read it as “who is junior to this date, and by how much,” not as a prediction of what the Division Engineer would actually order.

Pick a calling right

Real decrees from this basin, ordered by administration number — that is, by actual seniority. The years will look out of order, and that is correct: an administration number encodes the adjudication date as well as the appropriation date, so a right appropriated in 1970 can be administered ahead of one from 1950. A senior call curtails almost everything; a junior one curtails almost nothing.

Or enter an administration number directly

Administration number

DWR administration numbers encode appropriation and adjudication dates; lower is more senior.