secchi: Lake Tahoe sensor data from TEON, with context (provisional)

secchi data, through 2026-11-02 Open data behind https://brooksgroves.com/secchi/, built from the project's store by pixi run export. All data is provisional. Nothing here is an official TEON, USGS or other agency product; cite the upstream sources (SOURCES.md), not this bundle. Disclaimer These data are provisional. They are raw or lightly processed readings from instruments in the field and are subject to instrument performance, maintenance cycles and field conditions. They have not been reviewed or approved by TEON or any other agency, and their sources may revise, correct or withdraw them at any time. Re-download before relying on a number. This is not an official product of the Tahoe Environmental Observatory Network, the Tahoe Institute for Global Sustainability, the University of Nevada, Reno, the U.S. Forest Service, the Tahoe Science Advisory Council, the U.S. Geological Survey, USDA NRCS, NOAA, the National Weather Service or UC Davis TERC. No endorsement by any of them is stated or implied. The quality flags are not quality control. They are inferred by secchi from the data. A reading without a flag has not been certified; it is only one nothing here has caught. Flagged readings are kept exactly as stored. Provided as is, without warranty of any kind, express or implied, including accuracy, completeness, timeliness or fitness for a particular purpose. Do not use these data for safety-critical, regulatory, legal or operational decisions. Anyone relying on them does so at their own risk. Terms. Each source's own terms govern its data (see SOURCES.md). The MIT licence on the secchi code does not apply to the data. Cite the upstream sources, not this bundle. Files | dataset | rows | sites | first | last | what it is | |---|---:|---:|---|---|---| | lake_exo | 1,407,410 | 4 | 2025-03-20 | 2026-11-02 | Lake EXO sondes: telemetered (4H Camp, Glenbrook, Sunnyside) and hand-collected (Blackwood 3, Meeks). | | lake_nearshore | 1,875,790 | 6 | 2025-01-16 | 2026-09-28 | Nearshore self-logging MiniDOT (oxygen, temperature) and HOBO (temperature) loggers, read by boat. | | forest | 8,145,203 | 7 | 2024-06-13 | 2026-11-02 | Forest stations: soil, air, trees, stream level and chemistry, precipitation gauge. Soil_VWC is a fraction (0.034 = 3.4 %). | | usgs | 164,230 | 10 | 2026-09-16 | 2026-11-02 | USGS stream gauges in the Tahoe basin. | | snotel | 17,653 | 4 | 2024-06-01 | 2026-11-01 | NRCS SNOTEL: Ward Creek #3, Rubicon #2, Marlette Lake and the Snow Lab station. | | smoke | 2,652 | 1 | 2024-06-01 | 2026-11-01 | NOAA Hazard Mapping System smoke over the lake, daily. | | asos | 6,952 | 2 | 2024-06-01 | 2026-11-01 | National Weather Service daily weather at South Lake Tahoe and Truckee airports. | Each dataset is <name>.parquet (everything) and <name>_<year>.csv.gz (one per calendar year, for Excel and quick looks). SHA256SUMS checks every file; manifest.json lists row counts, columns and date ranges. Read it import duckdb duckdb.sql("SELECT site, avg(value) FROM 'https://github.com/bdgroves/secchi/releases/download/data-2026-11-02/lake_exo.parquet' WHERE variable = 'Temp' AND quality_flag IS NULL GROUP BY site") import pandas as pd df = pd.read_parquet('https://github.com/bdgroves/secchi/releases/download/data-2026-11-02/lake_exo.parquet') library(arrow) df <- read_parquet('https://github.com/bdgroves/secchi/releases/download/data-2026-11-02/lake_exo.parquet') Columns (TEON datasets) uuid record id from TEON, shared by the variables of one record; source; site; sensor_type; timestamp; lat, lng; variable; value; quality_flag. One row per reading per variable (long format). Timestamps are the loggers' clocks, with no zone written: Pacific Standard Time (UTC-8) all year, even in summer, except the MiniDOT loggers in lake_nearshore, which follow Pacific daylight time. USGS times carry their own offset. Soil_VWC is a fraction (0.034 = 3.4 %). Four naming conventions coexist: EXO Temp/Do_mgL/Chl_a, MiniDOT Temperature/Dissolved Oxygen, HOBO temperature, Campbell loggers Air_Temp/Soil_VWC/BattV_Avg. Units are not stored per row. See https://github.com/bdgroves/secchi/blob/main/docs/data-dictionary.md. Quality flags Values are exactly as stored. Nothing was corrected or removed. quality_flag is empty for readings with nothing noted, or one or more of these separated by ;. They are inferred by secchi from the data and are not TEON's own quality control. | flag | meaning | |---|---| | impossible_range (10,346 rows) | Value outside what the instrument could physically read in Lake Tahoe (water temperature outside -2 to 35 C, oxygen above 20 mg/L, oxygen saturation outside 50-150 %, pH outside 0-14). The same rules the watcher uses. | | channel_scramble (51,830 rows) | At this exact timestamp the sonde's channels were filed under the wrong names: oxygen concentration above 20 mg/L (saturation in the concentration field) flags the optical channels; water temperature above 35 C (saturation in the temperature field) flags every channel except battery. Inferred by secchi, not stated by TEON. The readings are probably intact and mislabelled, but are not remapped here. | | exo_sat_sea_level (120,000 rows) | EXO oxygen saturation (Do_percent). secchi's test (docs/dissolved-oxygen.md) found the EXO sondes reference saturation to sea-level pressure, so at Tahoe's altitude it reads about 20 points below the true saturation. Do_mgL is a measured concentration and is not affected. Reported to TEON; not confirmed by them. | | negative_turbidity (84,075 rows) | EXO turbidity below zero: not a real measurement, a zero-point offset. Common, not rare: every EXO sonde in the store reads below zero for much of its record, and at Sunnyside and 4H Camp the offset changes at service visits. Treat turbidity as relative until the zero point is understood. | | negative_chlorophyll (59,696 rows) | EXO chlorophyll (Chl_a) below zero: a zero-point offset, not a real concentration. Every EXO sonde does it, Sunnyside most (83 % of its record). Readings near zero mean very little chlorophyll, but the exact level isn't known; treat it as relative. | | negative_phycocyanin (25,887 rows) | EXO phycocyanin (blue-green algae pigment) below zero: a zero-point offset, as for chlorophyll and turbidity. Every EXO sonde does it some of the time (12-36 % of readings). | | ph_zero (30,104 rows) | EXO pH exactly 0: a dead channel. pH works only on the hand-collected sondes. | | repeated_hour (192 rows) | MiniDOT clocks follow daylight saving, so when clocks fall back (01:00-01:59 on the first Sunday of November) the same timestamps happen twice: two real readings an hour apart share each one. Both are kept; nothing in the record says which came first. | | duplicate_reading | Another row has the same site, sensor, variable and timestamp, under a different record id, with the same value. Keep one: QUALIFY row_number() OVER (PARTITION BY site, sensor_type, variable, timestamp ORDER BY uuid) = 1 | | conflicting_duplicate (146 rows) | As duplicate_reading, but the values differ. The cause has not been established (some are two readings within the same minute); not safe to drop either without looking. | Flags exist only on the TEON datasets. A reading with no flag is not certified good; it is just one nothing here has caught. Sites TEON hides This bundle leaves out any site TEON's /site-visibility/disabled list hides (checked: live; 1 hidden). Cite See CITATION.cff for this bundle and SOURCES.md for the upstream sources, which come first.