Wavelet analysis of the annual discharge records of the world’s largest rivers
Introduction
Interannual, decadal and multidecadal fluctuations are well documented in global and regional climatology. These oscillations have been emphasized by analysis of atmospheric disturbances such as sea level pressure time series, sea surface temperature (SST) fluctuations or continental surface temperature fluctuations. Indeed, hydrological signals can also bring consistent informations about climate fluctuations above the continental area.
Two well-known signals in hydrology can be of interest: precipitation and discharge records. Precipitation constitutes a noisier climate signal than discharge records. Indeed, discharge records are more easily measurable in general but small watersheds hydrological response only reflects local climatology.
To overcome these difficulties, we propose to focus on large watersheds discharge at the mouth. Indeed, large watershed response integrates and smoothes a large number of parameters such as rainfall, evaporation, topography, lithology, soil, and vegetation cover heterogeneities. However, the estimation of cyclic processes in runoff suffers several problems. The main lies in the relative shortness of the available time series, particularly in developing countries. Furthermore, regional human impacts can induce changes that cannot be easily separated from the natural trends. These effects generally affect monthly or annual fluctuations but do not affect multi-annual climatologic components. However, anthropogenic effects detected in discharge time series are prevently deleted and not taken into account in the present study. Labat et al. [22], [25] previously show that large river runoff records can be used to correlate the intensification of the water cycle to global warming at continental and global scale [18]. Wavelet analysis of large river runoff records demonstrates that hydrological signals also integrate and reflect interannual, decadal and multidecadal climate oscillations.
Oscillations put in evidence in large river discharge records are then tentatively correlated to climate proxies such as sea surface temperature fluctuations or sea level pressure fluctuations over Atlantic and Pacific since these climatic proxies considerably influence the total amount of water available over continental area. However, these relationships should be considered only as preliminary investigations and relationships should later on be further investigated through wavelet coherence analysis [56], [24], [36].
Section snippets
Large river runoff data
Most of the discharge data are available from the Global Runoff Data Center (GRDC) and Unesco River Discharge Database. Amazone discharges are provided by the Hybam Database (ANA, UNB, and IRD). Periods when impact of dams or irrigation on the natural hydrological regime of rivers are identified have been systematically removed from the original time series [22]. Selected rivers drain large basins (above 104 km2), gauging stations are located as near as possible to the outlet of the river and
Overview of wavelet analysis
Earlier applications of wavelets to discharge time series already allows to detect climatic oscillation [21], [23], [27]. The reader is referred to Ghil et al. [12] and Labat et al. [25] for a more complete bibliography on the applications of wavelets in geosciences.
Results and interpretation
Relationships between oscillations detected in the discharge records of the world’s largest rivers and climatic regional proxies that control the amount of water available over the basin are now exposed and discussed for each continent. Considered climate proxies and oscillations already put in evidence are concisely summarized in Table 1 together with used methods. Then, for each continent, we compare oscillations detected in this study (summarized in Table 1 and Fig. 1, Fig. 2) with
North America rivers discharge variability
North America river discharge records exhibit three modes of variability: a 5–8-year, a 10–13-year and a multidecadal 25–30-year fluctuation.
The 5–8-year oscillation appears as restricted over the southern part and over the central part of the continent and remains scarce in the northern part in accordance both with the interannual scale correlation put in evidence by Sun and Furbish [54] between river discharges of Florida and ENSO and with an interannual mode related to ENSO forcing in
Conclusion
In this contribution, long-term annual fluctuations of large river discharges were analysed during the last century using wavelet transforms. Many modes of climatic variability have been identified and they are most clearly intertwined. As demonstrated in this study, large rivers runoff exhibit ENSO-related and NAO-related interannual variability, decadal variability, 15–20-year and 28–30-year variability. Moreover, the longest series available also show multidecadal 40–50-year fluctuations.
Our
References (66)
- et al.
ENSO and the natural variability in the flow of tropical rivers
J Hydrol
(1997) Evidence for intensification of the global water cycle: review and synthesis
J Hydrol
(2006)- et al.
Rainfall-runoff relations for karstic springs – Part II: continuous wavelet and discrete orthogonal multiresolution analyses
J Hydrol
(2000) - et al.
Evidence for global runoff increase related to climate warming
Adv Water Resources
(2004) - et al.
Reply to comment of Legates et al
Adv Water Resources
(2005) Recent advances in wavelet analyses: Part 1. A review of concepts
J Hydrol
(2005)Oscillations in land surface hydrological cycle
Earth Planet Sci Lett
(2006)- et al.
Relationship between interdecadal fluctuations in annual rainfall amount and annual rainfall trend in a southern mid-latitudes region of Argentina
Atmos Res
(1999) - et al.
Spatial organization of decadal and bidecadal rainfall fluctuations in southern North America and Southern South America
Atmos Res
(2001) Spatio-temporal structure of the pentadecadal variability over the North Pacific
Prog Oceanogr
(2000)