Hydrology [H]

H24B MCC:3005 Tuesday 1600h

Impact of Interdecadal Climate Variability on Hydrologic Variables and Water Resources Management II

Presiding:J A Dracup, Department of Civil and Environmental Engineering, University of California, Berkeley; F Chiew, Department of Civil and Environmental Engineering, University of California

H24B-01 INVITED 16:00h

An overview of Pacific Decadal (climate) Variability impacts on hydroclimate and water resources management in the western US

* Mantua, N J (nmantua@u.washington.edu) , Climate Impacts Group, Box 354235 University of Washington, Seattle, WA 98195-4235 United States

Hydroclimate variations at decadal to interdecadal time scales have proven to be especially problematic for ecosystems and societies by causing multi-year periods of abundance or scarcity in key natural resources. A review of 20th century Pacific Decadal climate variability (PDV) impacts on the hydroclimate of North America highlights watersheds that have been especially sensitive to periods of multi-year abundance and deficits in water resources. PDV causes multi-year variations in the natural surface water supply, and it has had profound impacts on water resources in some of the major river basins in the western US. PDV impacts on water resources were not well-recognized by many hydrologists or water resources planners until late in the 20th century, yet in most western US watersheds allocation agreements and water resources projects had long been completed before this time. The existence of PDV has periodically amplified or reduced conflicts over fully allocated surface water supplies in the western US, and is likely to continue doing so in the future, so bringing the latest understanding of PDV and its impacts on hydroclimate to the water resources planning and management communities has great potential for societal benefits.

http://www.cses.washington.edu/cig

H24B-02 INVITED 16:15h

Earlier snowmelt streamflow in western North America--a response to Pacific decadal variability or climate change?

* Cayan, D R (dcayan@ucsd.edu) , Scripps Institution of Oceanography, 9500 Gilman Drive, La Jolla, CA 92093-0224 United States
* Cayan, D R (dcayan@ucsd.edu) , US Geological Survey, 9500 Gilman Drive, La Jolla, CA 92093-0224 United States
Stewart, I T (istewart@meteora.ucsd.edu) , Scripps Institution of Oceanography, 9500 Gilman Drive, La Jolla, CA 92093-0224 United States
Dettinger, M D (mddettin@usgs.gov) , US Geological Survey, 9500 Gilman Drive, La Jolla, CA 92093-0224 United States

Widespread trends toward earlier onsets of spring snowmelt and streamflow have occurred during the last five decades in snowmelt driven watersheds across western North America. Advances in spring timing of a few days to over three weeks are observed from a network of U.S. and Canadian stream gages. Such changes, if they continue, could pose an enormous challenge to water managers in providing a reliable water supply to the region. These changes are clearly a response to warmer winter and particularly warmer spring temperatures over the region, but a deeper question concerns the origin of the warming--is it simply a reflection of decadal climate fluctuations over the Pacific/North America sector or is it a reaction to broader scale climate warming? Evidence is presented suggesting that only part of the warming and only part of the streamflow changes can be attributed to fluctuations in the Pacific Decadal Oscillation (PDO). An equal or greater fraction of the variance appears to correspond with changes in global temperature over the period. Notably, the advances in streamflow timing occured during both cool (1948-1976) and warm (1977-1998) phases of the PDO. Further, although the PDO apparently reverted to its cool phase in 1999, spring temperatures in the West was anomalously warm and spring snowmelt-driven streamflow timing remained quite early during the last 6 years.

H24B-03 INVITED 16:30h

Australian hydro-climatic variability: Causes and Implications

* Franks, S W (stewart.franks@newcastle.edu.au) , University of Newcastle, School of Engineering, Callaghan, NSW 2308 Australia

Australia has long been known to experience marked climate variability on timescales ranging from inter-annual through to multi-decadal. Recent research is beginning to elucidate the climatic mechanisms by which different modes of hydrologic variability occur. Whilst inter-annual variability can largely be ascribed to the El Nino-Southern Oscillation, longer-term modes of variability are evident that correspond to the warming and cooling epochs evidenced in both Pacific and global sea surface temperature records. In this presentation, a review of the different climate modes that affect Australia will be provided within the context of the implications for hydrological and water resource design and management. Emphasis is also placed on the explanation of the different regional impacts of climate variability. This review demonstrates the key importance of understanding the longer-term modes of variability given contemporary hydrological risk estimation practices. The presentation concludes with a discussion of the possible origins of the multi-decadal variability, including an apparent link to multi-decadal solar variability. The implications for future climate risk estimation and the predictability of a `greenhouse-enhanced' future are then discussed.

H24B-04 16:45h

Impact of Interdecadal Hydroclimatic Variability on the Management of Water Supply for Melbourne, Australia

Chiew, F H (f.chiew@civenv.unimelb.edu.au) , Department of Civil and Environmental Engineering, University of Melbourne, Melbourne, VIC 3010 Australia
* Rhodes, B (bruce.rhodes@melbournewater.com.au) , Melbourne Water, GPO Box 4342, Melbourne, VIC 3001 Australia
Peel, M C (m.peel@civenv.unimelb.edu.au) , Department of Civil and Environmental Engineering, University of Melbourne, Melbourne, VIC 3010 Australia
Amirthanathan, G E (amir@civenv.unimelb.edu.au) , Department of Civil and Environmental Engineering, University of Melbourne, Melbourne, VIC 3010 Australia

This paper presents the implications of interdecadal hydroclimatic variability on the management of Melbourne's water supply system. Melbourne is the second largest city in Australia with a population of over three million. Streamflow and reservoir inflow data from seven water supply catchments and rainfall data from 18 stations are used in this study. The empirical mode decomposition (EMD) method is used to identify cycles in the long rainfall and flow time series. The EMD analysis is also applied to bootstrap samples from the original time series to test the statistical significance of the identified cycles. In EMD analysis, a time series is decomposed into a set of intrinsic mode functions that are mutually independent. The decomposition is based on the direct extraction of energy (variance) associated with various intrinsic time scales that are automatically and adaptively selected from the time series. The EMD is a relatively new technique that is able to deal with both non-linear and non-stationary data, and has several advantages over other spectral analysis techniques. The EMD analyses of data from Melbourne's water supply catchments show statistically significant interdecadal cycles in many, but not all, of the time series data. The study also shows that the rainfall, runoff and storage characteristics are different in different interdecadal periods. The region is considerably wetter, with the storages spilling more often (or less drawn down) during the negative IPO phase compared to the positive IPO phase. The El Ni\~no/Southern Oscillation versus hydroclimate teleconnection is also stronger during the negative IPO phase. The Interdecadal Pacific Oscillation (IPO) represents a low frequency ocean-atmosphere fluctuation in the Pacific region. However, the wet and dry cycles in the rainfall and flow data, and the cycles identified in the EMD analysis, are not necessarily in phase with the IPO periods. Nevertheless, although natural cycles are evident in the historical data, their future characteristics are difficult to predict. In addition, the future hydroclimatic characteristics are likely to be modified by global warming. This paper will discuss the implications of interdecadal hydroclimatic variability and climate change on the concept of reliability/security of water supply and the management of water resources.

H24B-05 17:00h

Low frequency drought variability in the western US from paleoclimatic indicators and GCM projections

* Hidalgo, H G (hhidalgo@ucsd.edu) , Scripps Institution of Oceanography, University of California, San Diego 9500 Gilman Drive, MC 0224, La Jolla, CA 92093-0224 United States
Cayan, D R (cayan@desert.ucsd.edu) , Scripps Institution of Oceanography, University of California, San Diego 9500 Gilman Drive, MC 0224, La Jolla, CA 92093-0224 United States

Low-frequency hydrologic variations of the western US for the past 500 years contained in gridded tree-ring reconstructions of Palmer Drought Severity Index (PDSI), were compared to PDSI projections under climate change scenarios calculated from the output of general circulation models (GCMs). Tree-ring results suggest that bidecadal and pentadecadal PDSI oscillations have been a common feature of the climate of the western US at least for the past 500 years. These variations are though to be related to similar low-frequency climate variations from the Pacific and Atlantic Ocean basins. Future PDSI projections computed from the GCM estimates of precipitation and temperature, also contain significant multidecadal variations, as well as significant negative trends. Although precipitation exhibits very little trend in many of the GCM projections for the western US, warming temperatures will drive PDSI into the dry category more frequently in the future according to the GCM data. Regional structure and change of precipitation is inconsistent across models. However, warmer climate alone will produce more frequent scarcity of water availability

H24B-06 17:15h

An exploratory study of seasonal rainfall variability in Australia using Independent Component Analysis

Westra, S (seth@civeng.unsw.edu.au) , School of Civil and Environmental Engineering, The University of New South Wales, Sydney, NSW 2052 Australia
* Sharma, A (a.sharma@unsw.edu.au) , School of Civil and Environmental Engineering, The University of New South Wales, Sydney, NSW 2052 Australia

Component extraction techniques have been used frequently by climate and water resources researchers to analyse high dimensional data sets. In this study we explore the application of a relatively new technique known as independent component analysis (ICA) to time series of Australian rainfall, as an alternative to the better known principal components analysis (PCA). The primary distinction between these two techniques is that while PCA uses only second order statistics to obtain uncorrelated components, ICA attempts to maximise independence between the components through the use of higher order statistics. Using a synthetic study that is designed to highlight important characteristics of Australian rainfall time series, we show that while PCA is better suited to the tasks of variance maximisation and dimension reduction, ICA is fundamentally more suited to ensuring the statistical independence of the extracted components and in certain cases is also capable of determining the underlying causes of this variability. As a result, we consider the ICA methodology to be more suitable for formulating a statistical basis for predicting rainfall, since the independence criterion allows for the formulation of distinct statistical models for each identified component. The sensitivity of the ICA to several factors including record length, the number of components extracted, the amount of noise in the data and the need for dimension reduction before the ICA, are examined in detail. The ICA technique is then applied to seasonal rainfall time series from over 200 rainfall gauges located around the Australian continent. The physical interpretability of the extracted ICs is assessed based on existing knowledge on the underlying causes of rainfall variability in Australia, and conclusions on the suitability of the technique for statistical forecasting are drawn.

H24B-07 17:30h

Observed trends in the hydroclimatology of Lake Superior from 1948 to 1999: A weakening seasonal cycle

* lenters, J D (jlenters@lssu.edu) , Lake Superior State University, School of Environmental and Physical Sciences 650 W. Easterday Ave., Sault Sainte Marie, MI 49783 United States

Water levels in the Laurentian Great Lakes experience a fairly pronounced seasonal cycle as a result of seasonal variations in runoff, evaporation, and other water budget components. Significant changes have been observed in these seasonal lake level cycles since 1860, particularly for Lakes Erie and Ontario, which are rising and falling earlier in the year. A recent analysis of the Lake Superior water budget reveals that this largest of the Great Lakes has also undergone significant changes since 1948. In particular, the seasonal rising and falling of Lake Superior has weakened in amplitude by approximately 20 percent from 1948 to 1999. Changes of this magnitude for such a large lake reflect significant 51-year trends in the monthly mean flux of water into and out of the lake (up to 1360 m3/s). Much of these changes can be attributed to trends in runoff, precipitation, and evaporation, indicating that the regulation of Lake Superior water levels has played a minimal role in the observed changes. An analysis of trends in regional air and water temperatures indicates that regional warming, particularly in the spring and summer, may help to explain some of the observed changes in the seasonal hydrologic cycle.

H24B-08 17:45h

Hilbert-Empirical Mode Decomposition (EMD) analysis of the Pacific Decadal Oscillation (PDO) and western U.S streamflow

Dracup, J A (dracup@ce.berkeley.edu) , Department of Civil & Environmental Engineering, University of California, Berkeley, 760 Davis Hall, Dept. of Civil & Environmental Engineering, Univerity of California, Berkeley, CA 94720 United States
* Leonardson, R (rebeccal@ce.berkeley.edu) , Department of Civil & Environmental Engineering, University of California, Berkeley, 760 Davis Hall, Dept. of Civil & Environmental Engineering, Univerity of California, Berkeley, CA 94720 United States

The Pacific Decadal Oscillation (PDO) Index and streamflow in the United States are studied to find temporal patterns on a range of scales. The Hilbert-EMD technique is used on monthly PDO data and on reconstructed uninhibited daily flow data. Hilbert-EMD analysis uses Empirical Mode Decomposition (EMD) to separate the data into sub-signals of various frequency ranges (such as weekly, inter-annual, and decadal), and a residue displaying a long-term trend. The modes are then analyzed using the Hilbert Transform, to determine power and time characteristics of the frequency sub-signals. Hilbert-EMD analysis has been used to study a wide variety of signals from within the geophysical record, including climate variability, earthquakes, and ocean waves. This project displays the temporal patterns of the PDO Index and streamflow at varying scales, and allows correlation analysis of individual frequency modes between the PDO and a number of major US rivers.