Methodology & sources

The City of Durham regularly sources drinking water from both Lake Michie (Flat River watershed) and Little River Reservoir (Little River watershed). This dashboard presents each reservoir independently and offers a combined view of monitored inflow.

• Lake Michie inflow — USGS 02085500, Flat River at Bahama, NC (discharge, 00060): monitored Flat River inflow toward Lake Michie.
• Lake Michie outflow — USGS 02086500, Flat River at Dam near Bahama (discharge, 00060): flow immediately downstream of the dam.
• Little River inflow — the sum of two monitored tributaries feeding Little River Reservoir: USGS 0208521324, Little River at SR1461 near Orange Factory, and USGS 0208524090, Mountain Creek at SR1617 near Bahama (both discharge, 00060). The two gauges report on the same schedule and are added together on each timestamp for a combined inflow figure. This is still only monitored inflow — other ungauged tributaries, direct rainfall, and runoff are not captured. The historical "typical for this date" context sums both gauges' day-of-year percentile bands; because percentiles are not strictly additive, the summed median is a sound estimate of typical combined flow while the surrounding bands are an approximation.
• Little River downstream flow — USGS 0208524975, Little River below Little River Tributary at Fairntosh, NC (discharge, 00060), about half a mile downstream of the dam. Because it sits below a tributary confluence, its readings include tributary flow — not pure dam release — and USGS frequently suppresses them as backwater-affected ("Bkw"). During those spells USGS cache nodes can also fall out of sync — some briefly return the whole record as no-data, others keep republishing the last pre-suppression value at fresh timestamps. The dashboard treats the outflow as "not currently reported" whenever USGS flags it suppressed, the value has been unchanged for over a day, or the latest reading is more than six hours old; it then shows the last measured value as context (with the date it last changed) and excludes it from net-flow figures, rather than presenting a frozen number as live. Brief gaps still fall back to the most recent stored value with its real timestamp. For this reason, Little River's rising/falling status is never derived from inflow−outflow; it requires measured elevation. No USGS elevation telemetry exists for either reservoir (the at-dam sites 02086490 and 0208524845 hold only discrete water-quality samples).
• Reservoir elevations (both) — City of Durham Lake Levels page, updated roughly daily and labeled provisional.

USGS site 02086490 (Lake Michie at Dam) holds only discrete water-quality samples (most recent February 2024) and publishes no real-time elevation series; no equivalent USGS elevation site is active for Little River Reservoir. The City of Durham is the only published source of current elevation for both reservoirs.

A stream gauge upstream of a reservoir measures water moving toward it — not the reservoir's level, storage, or health. Reservoir direction (rising/stable/falling) is shown only when elevation history or a full inflow/outflow pair supports it, and is never inferred from inflow alone.

The overview page may sum the two monitored inflows as combined monitored inflow to Durham reservoirs. The sum is computed only when both readings are sufficiently fresh, and it is not total Durham inflow, total water availability, or net water gain: it excludes ungauged tributaries, direct rainfall, withdrawals, evaporation, and other water-balance components.

Inflow-to-inflow comparisons are valid (same measurement, same unit). But raw cfs is not normalized for watershed size or hydrology, so it does not indicate which reservoir is filling faster, holds more water, or is healthier. Percent full and net storage change are not computed for either reservoir.

Each inflow reading is compared with the USGS daily-statistics service: the distribution of flows recorded on the same calendar day across the gauge's full period of record (1926–present for the Flat River; 1988–present for the Little River). The stated percentile is a piecewise-linear interpolation between USGS-published percentile points (5th–95th); values beyond those points are reported as "at or below the 5th" / "at or above the 95th" rather than extrapolated.

Water leaving the system through municipal use comes from the City of Durham's Current Data page: daily demand (MGD), the 30-day average, and days of supply. Days of supply is the City's own storage-derived estimate at the 30-day average demand, split into easily accessible (premium) water, water below the intake structures, and emergency storage (Teer Quarry). When demand is compared with combined monitored inflow, remember demand is served from reservoir storage — not directly from the gauged rivers — and the comparison excludes ungauged inflows, rainfall, releases, and evaporation.

No percent-full figure is shown anywhere: the City does not publish one on a stable page, reservoir geometry makes a linear estimate from elevation misleading, and no authoritative elevation-to-storage table has been located. Days of supply and feet below full pool are the honest health metrics available.

"What would it take to refill?" converts the elevation deficit into a volume using each reservoir's full-pool surface area: Lake Michie 508.7 acres (USGS bathymetric survey, WRI 93-4031, 1992) and Little River Reservoir 549 acres (City of Durham Parks & Recreation). Because a reservoir narrows with depth, deficit × full-pool area is an upper bound on the true volume — sustained net inflow at the stated rate is sufficient to refill in the period, and the true requirement is somewhat less. These are not derived from a stage-storage curve (none is published); the figures are labeled as bounds, not measurements. "Net inflow" is water accumulating in the lake after downstream releases, municipal withdrawals, and evaporation. Note the Lake Michie survey dates to 1992; sedimentation (~38 acre-ft/yr then) has reduced capacity since, which does not affect the area-based bound but does affect any capacity-percentage claims — another reason none are made.

Rainfall comes from Weather Underground personal weather stations — hobbyist gauges, not official instruments, so accuracy and siting vary. Lake Michie uses two upstream stations (Rougemont and Timberlake); Little River Reservoir uses one near its dam (Treyburn). Daily totals reset at local midnight, so "today" is rain so far. Where a reservoir has more than one station, the panel shows the mean across those reporting, as a rough watershed estimate. The City of Durham also publishes month- and year-to-date watershed rainfall (shown on the overview); the two sources measure different things and will not match exactly. A weather-station outage never affects reservoir data. Set a WU_API_KEY environment variable to use your own free PWS key in a deployment.

Drought status comes from the U.S. Drought Monitor data services (NDMC / USDA / NOAA) for Durham County, where both reservoir dams sit — though the watersheds extend into Person and Orange counties. Categories run D0 (Abnormally Dry) through D4 (Exceptional Drought); percentages are cumulative ("share of the county in this category or worse"). The DSCI (0–500) is the sum of the five cumulative shares; the four-week DSCI change drives the improving/worsening label. Maps are released weekly, so this panel updates on that cadence.

Flows can be displayed in cubic feet per second (cfs, the USGS native unit) or gallons per second (1 cfs = 7.480519 gal/s). Volumes use million gallons (1 cfs ≈ 0.646317 million gallons per day). The underlying data is always cfs; conversion is display-only.

Where both gauges exist (Lake Michie), we show an estimated net surface flow = inflow − outflow. This is a partial water balance only: it excludes municipal withdrawals, ungauged tributaries, direct rainfall on the reservoir, evaporation, groundwater exchange, gauge travel-time, and channel storage. The lake can fall even when net surface flow is positive.

Reservoir geometry is nonlinear, so percent-full cannot be inferred from elevation. We show elevation and feet below full pool (341.00 ft msl for Lake Michie, 355.00 ft msl for Little River Reservoir, per the City). A storage percentage will only appear if published by an authoritative source or derived from a verified elevation-to-storage curve.

• Rolling averages are the mean of instantaneous readings within the trailing window (24 hours, 7 days) before the latest reading.
• Direction uses measured elevation change over 24 hours when available (threshold ±0.02 ft — a product choice, not a hydrological standard). This dashboard stores each daily City elevation reading; once consecutive daily snapshots exist, direction is computed from measured lake level. Until then — or whenever the latest reading is stale — Lake Michie direction falls back to measured surface flow, clearly labeled, and Little River direction remains unknown.
• Flow-to-volume: 1 cfs ≈ 0.646317 million gallons per day.

The City publishes elevation history only as chart images. To seed the elevation charts, daily values for 2026-06-22 through 2026-07-18 were digitized by eye from the City's "Previous 30 Days" chart (accuracy roughly ±0.3 ft), calibrated against three precisely known readings. These points are marked with a "digitized" qualifier, are shown on charts for context, can never overwrite a precisely captured value, and are excluded from all rising/falling trend calculations, which use only precise scraped or City-published figures.

USGS and City data are provisional and subject to revision. Each source shows the age of its latest observation and is marked current, delayed, stale, or unavailable. Failures never substitute zero, and a failure in one source or reservoir never blocks the others.