4.1 High-Frequency Analysis of Oxygen Data Along
the Main Seine Stem
Dissolved oxygen was long used as an indicator of quality for streams and rivers,
mainly based on threshold values, e.g. the well-known concentration of 4 mgO 2 L
À1
under which aquatic life slows down [68]. Classically, dissolved oxygen data are
also used to calibrate biogeochemical models [22, 41]. Since the 2010s, there has
been considerable development of numerical routines for direct quantification of
metabolism from high-frequency dissolved oxygen monitoring [69, 70], constituting
a functional indicator of river ecological functioning.
High-frequency monitoring of dissolved oxygen has been carried out since the
1990s by the SIAAP (Syndicat Interdépartemental pour l’Assainissement de
l’Agglomération Parisienne) given the MeSeine monitoring network (see Fig. 2).
Sensors were first based on polarographic electrodes (Evita Oxy 4150) [71] and
since the 2010s on optodes (LDO Hach Lange, LXV416.99.20001) [72], which
allows for more stable and robust measurements. High-frequency oxygen data are
available at four stations downstream of Paris and upstream and downstream of the
SAV wastewater treatment plant, for two periods: 2002–2004 and 2015–2017. Data
typically exhibited a marked seasonal cycle, with low oxygen concentrations (undersaturation) during summer. The seasonal dynamics is first controlled by the temperature dependence of oxygen solubility and second by the increase of microbial
activity with temperature [73]. From a minimum concentration around 2 mg L
À1
during 2002–2004, the situation has been significantly improved over the last
15 years. Oxygen concentrations are now above the 4 mg L
À1 threshold.
Sub-hourly oxygen measurements at plurennial scales show typical seasonal
patterns and circadian cycles allowing for the calculation of ecosystem metabolism
[69, 70, 74, 75]. The quantification of the daily metabolism at a monitoring station is
based on the analysis of oxygen circadian variations, from which values of ER
(ecosystem respiration, defined as negative), GPP (gross primary production) and
NEP (net ecosystem production; NEP ¼ GPP + ER) are deduced. Each term ranged
from 1 to 30 gO 2 m
À2 day
À1 following a seasonal cycle with the lowest values in
winter and the highest in summer, within a range similar to values found in isolated
eutrophic ponds [74] but slightly higher than those found in freshwaters for low and
high Strahler orders [15, 41, 69, 76, 77].
For the two periods and all the stations, NEP is on average negative in this
urbanised sector, indicating heterotrophic functioning, which is fully consistent with
previous estimates [41, 69, 78]. NEP seasonal variability is more intense during the
2002–2004 period than during the 2015–2017 period, due to the reduction in algal
development and its respiration and to a reduction of bacterial heterotrophic respiration following the decrease in organic matter emissions from WWTPs (Fig. 3).
The spatial and temporal dynamics of metabolism along the reach and between
the two periods studied are characterised in terms of average daily metabolism for
each entire period (Fig. 7). During 2002–2004, the absolute value of ER and GPP
increased continuously from upstream to downstream, roughly doubling metabolism
Ecological Functioning of the Seine River: From Long-Term Modelling. . .
203
Précédent

- 212/430

Suivant