Reading Lake Level Data

The Parameters

USGS gauges on lakes and reservoirs report level data under standardized parameter codes. There are more of them than there look to be: the same measurement, a lake’s water surface elevation, is filed under 7 different codes, because most of them name the vertical datum the reading is measured from and one does not. PlainLakes carries 9 codes in all, listed here as USGS names them:

Elevation and gage height can describe the same water surface with wildly different numbers, because they use different zero points. Storage answers a different question entirely: not how high the water sits, but how much of it there is. Where a gauge reports more than one parameter, PlainLakes shows each series separately with its own unit and its own statistics.

Where a gauge files the same lake’s surface under more than one of the elevation codes, the reading shown is the one whose datum USGS names, in preference to the one code that leaves it unstated — a named reference frame is the only thing that makes an absolute figure mean anything beyond its own gauge. USGS also files quantities this site does not carry, and each is excluded for a stated reason rather than passed over: 72020, 72036, 72022, 99065, 63160, 72019. Two of them report reservoir storage in units other than acre-feet and one reports a depth measured downward rather than an elevation, so none can share a column with the readings above.

Datum Caveats: Why Absolute Numbers Mislead

An elevation is always "above something," and the something varies. Some gauges reference the National Geodetic Vertical Datum of 1929 (NGVD 29), others the North American Vertical Datum of 1988 (NAVD 88), and some reservoirs use a project-specific local datum defined by the dam operator. The same physical water surface can differ by a foot or more between datums, and gage height zeros are arbitrary site choices.

Across the sites PlainLakes ingests, USGS quotes site altitudes against five named reference frames (NAVD 88, NGVD 29, the International Great Lakes Datum, a local mean sea level determined at the site, and a canal-system datum) and files none at all for 258 of them. So "a fixed vertical datum" is not something a lake level can be quoted against in general: the frame has to be named per site, which is why every level reading here states the one it was measured from.

Where the parameter name and the reading disagree, PlainLakes says so. USGS files a reservoir's stage under the same water-surface-elevation parameter whether it is measured from a national datum or from the operator's own zero. On eight of the series here the reading sits hundreds or thousands of feet below the land-surface altitude USGS records for the very gauge reporting it. Topaz Lake, Nevada publishes about 91 ft from a gauge NWIS places at 4,900 ft, which is a local project datum, not an elevation. Those rows are labelled "Elevation (local datum)" wherever they appear, and their record pages say what the figure may and may not be compared with. The change and the trailing-year range stay valid throughout: a 7 ft drop is 7 ft of water from any zero.

• Never compare absolute levels across lakes. Different datums and different zeros make cross-site comparisons meaningless.
• Compare a lake to itself. The trailing minimum, maximum, and mean for the same gauge, parameter and daily statistic share one baseline, so position within that range is meaningful.
• Watch for datum notes when using other sources. A figure quoted from a dam operator may sit on a different datum than the USGS series for the same reservoir.

Daily Snapshots, Not Real-Time

USGS publishes two kinds of series for most gauges: instantaneous values, sampled as often as every 15 minutes, and daily values, one figure per day. PlainLakes loads the daily-value series, not the instantaneous one. That choice trades immediacy for stability: daily values are the series USGS curates for record-keeping, and they are the right resolution for reading trends over weeks and months.

Which daily figure: mean, extreme, or a single reading

A daily value is not "the level that day." It is one statistic computed over that day's readings, and USGS records which one alongside the series. Most are a daily mean, but a gauge may instead report the day's maximum, the day's minimum, or a single instantaneous reading taken at a fixed clock time. The distinction matters when you read a range: the 365-day low of a daily-minimum series is the lowest daily low of the year, which is not the same quantity as the lowest daily average.

Of the 127 gauge series behind lake pages here, 76 are a daily mean and 51 are not. Every lake page names the statistic beside its reading, and where the series is not a mean it says so in words. For 7 of them USGS files more than one daily statistic for the same gauge and parameter; PlainLakes prefers the daily mean, then a timed observation, and takes a maximum or a minimum only when nothing else is filed. That preference is PlainLakes' own and the page states which statistic it is showing.

The practical consequence: a level shown here describes a recent day, not this minute. For fast-moving situations, flood operations, rapid releases, or launching a boat on a fluctuating reservoir, go to the USGS real-time feed for the gauge. How the batch pipeline works end to end is described in the data collection guide and in the site methodology.

Seasonal Drawdown vs Drought

The single most common misreading of lake level data is treating every decline as drought. Managed reservoirs move by design:

• Flood-control drawdown. Pools are lowered ahead of the wet season to create storage space for incoming runoff. The "empty" look in late autumn is the system working.
• Irrigation and supply cycles. Storage builds through spring snowmelt and drains through the growing season, year after year.
• Winter drawdown. Northern lakes are often lowered before freeze-up to protect docks and shorelines from ice damage.
• Hydropower scheduling. Generation releases produce fluctuations unrelated to weather.

Drought reveals itself as a broken pattern rather than a low number: the spring refill that does not arrive, a summer level far below the same season's usual range, or a multi-year stair-step downward. Natural lakes without operating schedules track precipitation and evaporation more directly, so for them a sustained decline carries more climatic signal.