Dissolved oxygen (µmol/kg)
Dissolved oxygen (QC'd)
- 1,317,091
- values
- 1,286
- flagged
i
1,286 of these 1,317,091 values (under 0.1 %) are flagged by their provider as questionable or bad. The Explorer, the climatology and its anomalies leave them out; do the same when you read the data yourself (cc_qual_ok_sql()in calcofi4r,qual_ok_sql()in calcofi4py). Theoxygen_umol_kg_ave_sta_corrseries from CTD Cast Files is an average of two sensors, rebuilt with any sensor its provider flags questionable or bad left out, so it carries no flag of its own. - 35,642
- sampling events
- 1949–2025
- years
- 2
- datasets
- 0–5,351 m
- depth
- every month
- seasons
O2 · dioxygen · 31.998 g/mol · CHEBI:15379
Concentration of dissolved oxygen per unit mass of the water column. Oxygen may be expressed in terms of mass, volume or quantity of substance.
Automated Winkler titrator, UV end-point detector
350 nm
Bottle sampling methods on calcofi.org ↗How large are the ocean’s “dead zones” and how fast are they changing?
CalCOFI holds 1,317,091 values of it from 35,642 sampling events in 2 datasets, 1949–2025, from the surface to 5,351 m. At 200–500 m it has fallen 7.26 per decade since 1984.R Below the mixed layer off Southern California oxygen has fallen since the 1980s and the hypoxic boundary has risen toward the shelf, squeezing the habitat of fish and invertebrates.C
other ways to say why · 6
- Oxygen is the balance of what the ocean takes in at the surface and what animals and bacteria use below; a warmer ocean holds less and mixes less, so it is losing oxygen. needs a citation before it can be the pick
- Warming thins the oxygen the surface layer can hold, and the upper ocean has been losing it for decades across the North Pacific, not only here. needs a citation before it can be the pick
- GOOS asks How large are the ocean’s “dead zones” and how fast are they changing? source
- GOOS asks Is the biomass of the ocean changing? source
- Wikipedia Hypoxia (environmental) source
- Wikipedia Ocean deoxygenation is the reduction of the oxygen content in different parts of the ocean due to human activities. There are two areas where this occurs. Ocean deoxygenation · CC BY-SA 4.0 · rev. 1354922617
The measurement
Measured in i
One row per series, one cell per year, darker with more values — the same strip the front
door draws per dataset and a species page draws per taxon. Each row is one dataset's own
measurement_type; the label is written once, in the title above.
i
One row per series, one cell per year, darker with more values — the same strip the front door draws per dataset and a species page draws per taxon. Each row is one dataset's own measurement_type; the label is written once, in the title above.
-
series
oxygen_umol_kg· source columnoxy_umol_kg· flag columno_qual1,286 flagged · 0.2 % -
series
oxygen_umol_kg_ave_sta_corr· source columnox_aveu_m_sta_corrsensor mean no flag at this grainDO average station-corrected
The structure
RDKit 2026.03.6 from the ChEBI molfile, in the page’s ink · CC BY 4.0. ChEBI’s roles are never shown.
How it is measured
Carpenter's (1965) modification of Winkler's 1889 method: manganous hydroxide is oxidised by the sample's dissolved oxygen, acid frees iodine from iodide in proportion, thiosulfate titrates the iodine, and an automated titrator reads the end point by the UV absorbance of tri-iodide rather than the visible starch-iodine color. (calcofi.org Bottle sampling methods)
- Mn2+ + 2OH- -> Mn(OH)2
- 2Mn(OH)2 + O2 -> 2MnO(OH)2
- MnO(OH)2 + 4H+ + 2I- -> Mn2+ + I2 + 3H2O
- I2 titrated with thiosulfate
Why it matters
The familiar scale i
Bars are each series' 5th–95th percentile, the line its minimum to maximum, the ring its median;
the marks are metadata/measurement_scale.csv's, each hovering to its own source.
i
Bars are each series' 5th–95th percentile, the line its minimum to maximum, the ring its median;
the marks are metadata/measurement_scale.csv's, each hovering to its own source.
The record’s own history i
Deeper than the climatology reaches: 16,274 values at 1000–2000 m, 6,649 values at 2000+ m have no
normal to depart from, because the release’s climatology stops at its
deepest built bin. Carrying it to the bottom is one argument of
build_climatology() and puts these bands on the page too.
i
Deeper than the climatology reaches: 16,274 values at 1000–2000 m, 6,649 values at 2000+ m have no
normal to depart from, because the release’s climatology stops at its
deepest built bin. Carrying it to the bottom is one argument of
build_climatology() and puts these bands on the page too.
the GOOS specification sheet for this Essential Ocean Variable lists it, and asks
GOOS asks
- How large are the ocean’s “dead zones” and how fast are they changing?
- Is the biomass of the ocean changing?
GOOS specification sheet ↗ · quoted, not copied
The record
Ways in i
Read from measurements.json of release
v2026.09.11, written at release by
calcofi4db::build_measurements_catalog() beside
taxa.json. The catalog's index is
/measurements/; the JSON-LD on this page is a
schema.org/DefinedTerm in the NERC P01 term set with a
PropertyValue per series.
ExplorerExplorer · a depth section vs normaldb-queryi
Read from measurements.json of release v2026.09.11, written at release by calcofi4db::build_measurements_catalog() beside taxa.json. The catalog's index is /measurements/; the JSON-LD on this page is a schema.org/DefinedTerm in the NERC P01 term set with a PropertyValue per series.
SELECT * FROM read_json('…/v2026.09.11/catalog.json')
-- objects[] WHERE table = 'obs_env'
-- AND partition_value = 'oxygen_umol_kg'
-- | 'oxygen_umol_kg_ave_sta_corr'library(calcofi4r)
con <- cc_get_db() # the promoted release
tbl(con, "obs_env") |> filter(measurement_type %in% c("oxygen_umol_kg", "oxygen_umol_kg_ave_sta_corr"))import calcofi4py as cc
con = cc.cc_get_db()
con.sql("SELECT * FROM obs_env WHERE measurement_type IN ('oxygen_umol_kg', 'oxygen_umol_kg_ave_sta_corr')").df()By depth i
Values per depth band at the release grain, one thin bar per series. The bands are the
record's own (0-10 … 2000+ m).
i
Values per depth band at the release grain, one thin bar per series. The bands are the record's own (0-10 … 2000+ m).
By month i
Values per calendar month, one row per series: CalCOFI's quarterly cruises show as peaks in
the months the ships sail.
i
Values per calendar month, one row per series: CalCOFI's quarterly cruises show as peaks in the months the ships sail.
Range & quality
details
- bounds
- The registry declares
-20 … 700 µmol/kgforoxygen_umol_kg · oxygen_umol_kg_ave_sta_corr; a value outside it is dropped at the ingest (drop_out_of_bounds()). - observed
- Hydrographic Bottle
oxygen_umol_kg-0.4349 · median 151.5 · 485.7 µmol/kg (5th–95th percentile 15.65–264)
CTD Cast Filesoxygen_umol_kg_ave_sta_corr0 · median 107.3 · 691.4 µmol/kg (5th–95th percentile 13.95–262.2) - flags
- Hydrographic Bottle
oxygen_umol_kgo_qualon 29,722 of 690,697 rows:628,436 ·81,285 ·91; qual_ok keeps 689,411. The codes are this dataset's own vocabulary, uninterpreted.
CTD Cast Filesoxygen_umol_kg_ave_sta_corrcarries no quality code at the release grain; qual_ok keeps all 626,394. - baseline
- A monthly climatology exists for this measurement (the release's climatology table: 1993–2013, dataset × grid cell × calendar month × 10 m bin), so the Explorer can draw anomalies against it rather than values.
Related measurements i
The same quantity from another platform or another kind of sample, each its own page and
never merged into this one: a dataset's own bottle samples would count that dataset twice, an underway intake
is not a cast, and a replicate is not a mean. The reason is the record's
(related[].why).
i
The same quantity from another platform or another kind of sample, each its own page and never merged into this one: a dataset's own bottle samples would count that dataset twice, an underway intake is not a cast, and a replicate is not a mean. The reason is the record's (related[].why).
- Bottle dissolved oxygen oxygen_btl_umol_kg · CTD Cast Files · 1993–2025 · 127,192 values · P01 DOXMZZXX the same quantity from the CTD files' own bottle samples — plausibly the same physical bottles, so merging them would count the bottle dataset twice
- DO sensor 1 (uncorrected) oxygen_umol_kg_1 · CTD Cast Files · 1993–2026 · 671,055 values · P01 DOXMZZXX one sensor of the pair beside the mean the file publishes
- DO sensor 2 (uncorrected) oxygen_umol_kg_2 · CTD Cast Files · 1999–2026 · 328,625 values · P01 DOXMZZXX one sensor of the pair beside the mean the file publishes
- Reported oxygen (pre-QC) r_oxygen_umol_kg · Hydrographic Bottle · 1949–2021 · 690,698 values · P01 DOXMZZXX the source's own reported value before its QC — this page is the QC'd twin
Cite
Cite the release; each dataset's own citation, licence and DOI are on its page — Hydrographic Bottle, CTD Cast Files. The Explorer's Share → Cite writes it for the view you make.