Title | Analyse des changements de régimes dans les séries temporelles issues de la dendrochronologie |
Download | Download (whole publication) |
| |
Licence | Please note the adoption of the Open Government Licence - Canada
supersedes any previous licences. |
Author | Perreault, L; Nicault, A; Boucher, É; Arseneault, D; Gennaretti, F |
Source | Utilisation des archives naturelles pour la reconstitution du passé hydro-climatique; by Bégin, C; Nicault, A; Bégin, Y; Geological Survey of Canada, Open File 8768, 2021 p. 130-142, https://doi.org/10.4095/328084 Open Access |
Year | 2021 |
Publisher | Natural Resources Canada |
Document | open file |
Lang. | French |
Media | on-line; digital |
Related | This publication is contained in Utilisation des archives
naturelles pour la reconstitution du passé hydro-climatique |
File format | pdf |
Province | Quebec; Newfoundland and Labrador |
NTS | 12M; 13D; 13E; 13L; 13M; 22M; 22N; 22O; 22P; 23; 32M; 32N; 32O; 32P; 33 |
Area | Complexe La Grande; Rivière Caniapiscau; Churchill Falls; Labrador |
Lat/Long WENS | -79.5000 -62.0000 55.0000 51.0000 |
Subjects | environmental geology; hydrogeology; Nature and Environment; Science and Technology; climatology; paleoclimatology; hydrologic environment; dendrochronology; modelling; models; meteorology; temperature;
precipitation; watersheds; statistical methods; sampling methods; Le projet ARCHIVES; Picea mariana; Methodology; Climate change; Hydrology; Boreal ecosystems; Forests; Trees; Biology; Artificial intelligence; Phanerozoic; Cenozoic;
Quaternary |
Illustrations | time series; models; flow diagrams; tables; plots; sketch maps; photographs |
Program | Climate
Change Geoscience Extreme Events |
Released | 2021 06 28 |
Abstract | The problem of switching regimes in hydrological time series led to serious questioning in Québec during the last decades. In fact, when we examine annual inflow and precipitation time series of a
number of Québec watersheds, we can identify alternating sequences of high and low values. These variations must be taken into account in the hydrological forecasting process. However, since we only have limited information (no more than five decades
of observations), we may examine different natural archives such as trees to study hydroclimatic variability over a longer period. One of the objectives of the ARCHIVES project was therefore to study hydrometeorological regime changes in time series
reconstructed using dendrochronology. In this chapter, we address the problem of changepoint analysis in such time series by using finite mixtures of distributions. Mixtures of distributions become natural models to represent datasets in which
observations may originate from several distinct statistical populations. The problem is treated from a Bayesian perspective. Our approach was applied to several time series reconstructed in the ARCHIVES project. These applications allowed us to
identify regime changes that are spatially consistent, as well as historical changepoints that correspond to specific events such as abrupt changes in summer temperatures that coincide with strong volcanic eruptions. In this chapter, we describe the
approach and illustrate it by analysing time series of summer temperatures and water inflows reconstructed using black spruce (Picea mariana Mill. BSP) tree-rings. |
GEOSCAN ID | 328084 |
|
|