Sampling-date drift hid part of the deep-water oxygen loss in 183 temperate lakes (1941–2017), and in 120 northern lakes the loss was present by mid-June with no change in summer depletion rate
Jane et al. (2021, Nature) reported that dissolved oxygen (DO) decreased in temperate lakes. They used the mean of all profiles from 15 July to 31 August in each year. Deep-water DO decreases quickly inside this window, so a drift in sampling dates can change the trend. We used the public data (EDI edi.698.4) and did three tests. First, we reproduced the published medians: surface temperature +0.373 °C per decade (paper +0.39), surface DO -0.107 mg/L per decade (-0.11) and deep DO -0.111 mg/L per decade in 183 lakes (-0.12). Second, we measured the sampling-date drift. Deep DO decreased by 0.043 mg/L per day inside the window. In 23% of lakes, the mean sampling day moved by more than 3 days per decade. In deep-water lakes, sampling moved slightly earlier, and this hid part of the decline. With a within-year date correction, the median deep DO trend became -0.148 to -0.161 mg/L per decade (three methods). The paired change per lake was -0.022 mg/L per decade (95% CI -0.042 to -0.004). Surface trends did not change by a meaningful amount. Third, we fitted a seasonal model to a fixed-depth deep layer in 120 northern-hemisphere lakes. The DO trend at 15 June (-0.097 mg/L per decade, CI -0.219 to -0.054) was equal to the trend at 15 August (-0.095, CI -0.181 to -0.031). The change in the 30-day depletion was +0.001 mg/L per decade (CI -0.031 to +0.035). In 89 lakes, the deep-layer DO trend at 15 May was already -0.135 mg/L per decade. Thus, the loss occurs before or near the start of summer stratification and not from faster summer depletion. This agrees with Jane et al. (2023), and we used a method that needs no stratification model. Lakes with less clarity had a larger loss at 15 June (Spearman 0.41, n = 117), but clarity had no relation to the depletion-rate change. Limits: the lakes are mostly in North America and Europe and the series end in 2017-2018. The date correction is linear. The deep layer is a fixed-depth concentration and not a volume-weighted value. We did not measure the stratification onset date directly.
lake deoxygenation · dissolved oxygen · hypolimnion · thermal stratification · sampling date bias · seasonal decomposition · reproduction · long-term monitoring · water clarity · temperate lakes · climate change
References
6- doi:10.1038/s41586-021-03550-y
[Legacy relation: reproduces] We reproduce its median lake DO and temperature trends from the public data and test its late-summer window method for sampling-date drift.
- usesdoi:10.6073/pasta/7686da566d3589b95b01c530c8cd0a22
EDI package edi.698.4: the lake temperature, DO and Secchi profiles that we analyse (CC BY 4.0).
- supportsdoi:10.1111/gcb.16525
Our fixed-date decomposition from observed profiles independently agrees with its finding of no consistent change in seasonal DO depletion rates.
- doi:10.1016/j.envsoft.2011.05.006
Definition of metalimnion depths (rLakeAnalyzer) that we implement again in Python.
- doi:10.4319/lo.1992.37.6.1307
Oxygen solubility equation used for percent saturation.
- doi:10.1038/s41467-021-22657-4
Earlier onset and longer duration of lake stratification under warming, a possible cause of the early-season DO loss.
Cited by
6 publicationsPublic catalog entries on Sciscape.
- Removing the deepest oxygen reading reduced gradient flags from 20 to 4 in 123 Wisconsin lake-years selected for March–June profile temperature ranges of at most 1°C
This prior analysis reports early-season oxygen loss and motivates a separate profile-level test of a temperature rule.
- usesThe median oxygen trend gap between middle and lower fixed sampled depths was uncertain in 149 northern lake records, 1980–2017
Its sample-date result guides the date adjustment. We reuse unchanged source data from its retained dataset.
- supportsDeep-water oxygen loss is a shared decline across 195 temperate lakes, but lake-specific trends did not persist between 1990–2003 and 2004–2017 or predict 2018–2025
Our independent rebuild also finds that sampling-date correction and fixed-depth layers make the median deep-water trend slightly more negative (-0.14 to -0.15 mg/L per decade).
- supportsJune oxygen reduced August prediction error in fixed linear models for 40 northern lake records in 2001–2017
Its finding that oxygen loss was present by June motivates the later-year prediction test. No derived layer or fitted result is reused.
- usesCool, oxygen-rich depth had a mean within-region annual correlation of 0.23 in 96 northern lakes, 1980–2017
Its sampling-date result motivates our date term and close-date checks.
- usesWith a 5 mg/L oxygen limit, adding an 18°C limit raised the median paired depth slope by 0.80 percentage points per decade in 149 northern lakes, 1980–2017
We reuse its retained unchanged EDI source snapshot and address the sampling-date issue it identified. We wrote new depth accounting and analysis code.