Combining multi-variate palaeoecological indicators and mining closed gradual patterns for refining past lake dynamics and the induced ecological legacies
Résumé
The overloading of water bodies with nutrients is a widespread ecological alteration. A
recent synthesis (CNRS, IFREMER, INRA, IRSTEA) shows that lake ecosystems are partic-
ularly concerned by the eutrophication process which results in: (1) ecological disturbances
(e.g. invasive species, loss of biodiversity, algal and cyanobacterial blooms), and (2) pub-
lic health problems (e.g. water quality degradation due to cyanotoxin production). This
makes the restoration of eutrophic freshwaters an environmental top-priority and a key-issue
for scientic research. Accurate assessments of the present-day lake ecosystems state are
thus needed. Moreover, long-term retrospective models of lake dynamics must be developed
concomitantly to the analysis of natural and anthropogenic modications of the catchment,which enhance the external input of sediment and nutrients. Palaeoecological researches are
therefore required because they characterize through time the lake’s responses to cumula-
tive changes caused by natural (e.g. climate) and anthropogenic impacts (e.g. vegetation
clearance, agriculture). They furnish thus fresh insights into a good understanding of the
ecological legacies which also determine current lake structure and function, and which must
be considered for sustainable lake management. A palaeoecological research was conducted
in the hyper-eutrophic Lake Aydat (837 m a.s.l.), located in the Chaˆıne des Puys (French
Massif Central). The high resolution and multi-proxy analysis (pollen, non-pollen paly-
nomophs, diatoms, sedimentology) was combined to an approach of data mining (extraction
of frequent (closed) gradual patterns of multi-variate indicators under temporal constraint)
and allow to address the:
(1) reference conditions of the lake (prior to extensive human impact) and its natural vari-
ability;
(2) aquatic changes phases: timing, intensity, frequency, delay;
(3) reversibility or directions in which the lake is driven by the long-term cumulative impact
(e.g. loss of resilience, ratchet effect);
(4) potential drivers: climate
vs
the diverse range of land uses;
(5) degree of resistance/sensitivity and vulnerability of present-day lake ecosytems.