Introduction

Three PNW river basins are explored in the study: The Columbia Basin (at The Dalles), The Cedar and Rex Basins (Inflow to Chester Morse Lake), and the Yakima Basin (at the Parker Gage). Each of these basins is associated with a number of uses. The Columbia dam system is used primarily for flood control, hydropower production, irrigation, recreation, and maintenance of instream flow. Chester Morse Lake is one of two reservoirs supplying Seattle's Water. Demand is predominantly for M&I water supply, and the system storage is used to maintain instream flow and for flood control. The Yakima system dams are primarily used for flood control and to supply irrigation water.

The basins are also different in their hydrologic response. The Yakima is predominantly a snowmelt driven river, and shows the heavy emphasis on spring peak flows associated with this type of catchment. The Columbia is also heavily influenced by snow melt and experiences high peak flows in the spring, but there is a substantial component of the flow that comes from winter runoff in the lower parts of the basin that contributes to flow at The Dalles. Inflow to Chester Morse Lake, whose catchment is at moderate elevation on the Western slope of the Cascades, is split fairly evenly between winter runoff and spring snowmelt, and a typical hydrograph shows a winter and spring peak of comparable size.

Classification of Water Years

Twenty one classifications of water years are used. ENSO states are defined by the Dec-Feb NINO3.4 index. ENSO State 1(El Nino) is defined as those years for which the Dec-Feb average NINO3.4 index is more than 0.5 standard deviations above the mean. ENSO State 3 (La Nina) is defined as those years for which the Dec-Feb average NINO3.4 index is more than 0.5 standard deviations below the mean. The rest of the record is defined as ENSO State 2 (Neutral). Similarly all years are divided into PDO State 1 (more than 0.5 standard deviations below the mean for Nov-March average), PDO State 3 (more than 0.5 standard deviations above the mean for Nov-March average), and the remaining years are defined as PDO State 2 (Neutral). Epochal PDO signal is defined as either negative or positive. The epochal PDO signal is defined as negative for the periods from 1900-1924 and 1947-1976 and positive from 1925-1946 and 1977-present. During these periods, the PDO tended to remain predominantly in either positive or negative phase. ;On average, positive phase PDO years are associated with conditions that produce lower average streamflow in the basins than negative phase PDO years, and negative NINO3.4 years (La Nina) are associated with conditions that produce higher average streamflow than positive NINO3.4 years (El Nino).

The climate classifications that are combinations of the two climate indicators are used to determine the extent of any positive or negative reinforcement of effects associated with the two climate indicators. There are some problems with this approach since the sample size is different for each of the sub categories, and some of the categories have only a few years of data, which could potentially alter the results in a misleading manner if the sample size is too small. There are also potential multi-year effects in the basin's response to climate caused by impacts to system storage being carried over to the following year. If these multi-year effects frequently span water year climate classifications, then another type of confounding effect not evaluated here may occur. Despite these potential problems, the results clearly show the positive reinforcement that exists between the two climate indicators in the extremes (e.g., PDO negative/La Nina years and PDO positive/El Nino years).

Explanation of Graphics

The results of the study are contained in three MS Powerpoint files, each containing 24 graphs, three composite graphics showing average changes in hydrology for PDO and ENSO by month, and one scatter plot for each of the climate categories. An example of the scatter plot format is shown below. The example shows April-September average streamflow at The Dalles for El Nino water years (red triangles) compared to all Non-El Nino water years (blue circles). The two horizontal lines on the graph show the average of all El Nino and Non-El Nino years. For convenience in creating the graphs the legend was generalized to identify "category" and "non-category" results. "Category" in this case would be El Nino and "non-category" would be Non-El Nino. This convention is used throughout. In general the strength of the signal in terms of streamflow response can be quickly assessed by the separation between the two horizontal lines on the graph. If the lines are close together, the signal is weak (i.e. the category and non-category have roughly the same long-term average values), if they are far apart the signal is strong (i.e. the category and non-category have very different long-term average values). Variability and distribution of values can also be visually assessed using the scatter plots. It should be noted that some combinations of climate indicators are associated with only a few years in the record. In these cases some caution should be used in interpreting the results, since long term patterns may be masked by the limited data. This is particularly true of the Chester Morse Lake Inflow, which is estimated from a flow record from 1946-1994. Results from the Columbia and Yakima basins are generally better in this regard.