Link to CIG Homepage

Climate dynamics


Theme page links

Link to Integrated Assessment Theme Page Integrated Assessment
Link to Climate Dynamics Theme Page Climate Dynamics
Hydrology
Link to Salmon Theme Page Salmon
Link to Forests Theme Page Forests
Link to Coasts Theme Page Coasts
Link to Human Dimensions Theme Page Human Dimensions

Major Topics:

1. Overview

1.1 Physical Geography and Mean PNW Climate of the Twentieth Century

1.2 Patterns of Climate Variability in the Twentieth Century

1.3 The El Niño/Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO): Major Influences on PNW Climate

1.4 Implications for Climate and Resource Predictions

 

Links

Overview

Two guiding principles our integrated assessment of climate impacts on the Pacific Northwest (PNW) are: (1) that the temporal rhythms of climate variation in the region are driven by those in the larger climate system; and (2) that the interplay between hemispheric-scale climate patterns and the terrain of the PNW determines the spatial patterns of PNW climate variability.

Although the interactions between large-scale climate variations and the terrain of the PNW induce significant heterogeneity in microclimates and ecosystems within the region, there is a strong regional coherence in the variability of PNW climate. More simply put, "wet" winters tend to be wet throughout the 3-state region of Idaho, Oregon and Washington, as do "cold", "warm", and "dry" winters. There are exceptions to this general rule, but they have been relatively uncommon over the past century.

Research clearly documents the fact that very large scale "modes" of climate variability, in particular El Niño-Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO), exert important and sometimes predictable influences on PNW climate. We find that, given current climate prediction tools, it is the connections between the hemispheric and regional scales of climate variability that give rise to skillful predictions for PNW climate, at least at seasonal to inter-annual time scales. An additional important finding is that certain combinations of winter temperature and precipitation variations are amplified in the region's water cycle, and this sensitivity provides an avenue for extending climate predictions to predictions for water-sensitive natural resources like mountain snow pack and water year (October-September) streamflow.

 

 

1.1 Physical Geography and Mean PNW Climate of the Twentieth Century

The mean climate of the PNW region is strongly shaped by the interaction of seasonally varying pressure and wind patterns with PNW topography. Beginning about mid-October a semi-permanent low-pressure cell, commonly called the Aleutian Low (Fig.1), intensifies and migrates southeastward to a location centered over the Aleutian Islands and Gulf of Alaska. At seasonal-mean time scales, winter surface winds blow in a counterclockwise circulation around the Aleutian Low. To the south, winds blow in a clockwise circulation around a semi-permanent center of high pressure typically centered offshore southern California. Together, these high and low pressure cells typically bring moist, mild, onshore southwesterly and westerly flow into the PNW from October through early spring. Condensation occurs as the moisture-laden air rises and cools along the windward slopes of the mountains. The PNW's "wet-season" typically begins in October, peaks in mid-winter, and ends in the spring; about 75% of the region's annual precipitation falls in the period October-March. During late spring, the Aleutian Low retreats to the northwest and becomes less intense, while the high pressure cell to the south expands northward and intensifies. The result is a strong reduction, from late spring through summer, of onshore winds and precipitation bearing storms for the PNW.

The topographic map (Fig. 2) reveals the PNW's complex terrain. There can be enormous spatial variations over very short distances (less than a few km) in PNW climate because of landscape interactions with seasonal changes in wind and pressure patterns. The North-South oriented Cascade mountain range plays, for instance, a dominant role in the water resources and ecosystems of the region, effectively dividing the PNW into eastern and western sub-regions. The east-west contrast in annual mean precipitation (Fig.3) is dramatic. West of the Cascade crest, precipitation is generally in excess of 1.25 meters per year, while east of the Cascades annual precipitation totals less than 38 cm are the norm.

The west (i.e., windward) slopes of the Cascade and Olympic mountains receive enormous quantities of rain and snow, exceeding 5 meters (~200 inches) of water equivalent per year at some locations on the Olympic Peninsula. At Paradise Ranger station on Washington's Mount Rainier, the average spring sees an "end of season" snow depth of 4.1 meters. The Cascades are often among the snowiest places on Earth: in 1956 the snow at Paradise piled to a depth of nearly 9.1 (~28 feet) meters during a year in which that location's annual snowfall was 28.5 meters; the Mount Baker Ski Area, located in the north Cascades near the U.S./Canada border, set a new world record for the highest ever recorded annual snowfall (October-September) in 1998-99 with a total of 29 meters (~100 feet).

Although the west side of the Cascades is generally a very wet sub-region of the PNW, it contains several areas that receive significantly less precipitation than the west-side average. Washington's Puget Lowlands, the northeast extreme of the Olympic Peninsula, the San Juan Island archipelago, and Oregon's Willamette Valley are relatively dry areas that lie in "rain shadows." Rain shadows in these areas are caused by high terrain located to the west and southwest that shields them from the direct impact of storms that follow the wet-season's prevailing storm track.

The rain-shadow effect is especially strong for the region east of the Cascade crest. Most areas east of the Cascade range receiving more than 50 cm per year have west-facing slopes and relatively high elevations (e.g., northeast Oregon's Blue and Wallowa Mountains, Idaho's Bitterroot Mountains, and northwest Montana's Rocky Mountains).

The Cascade Mountains also bear strongly on seasonal variations in the region's climate because they present an effective barrier between the lower-atmosphere's maritime climate influences to the west and the continental climate influences to the east. West of the Cascades, the low-lying valleys have a maritime climate with typically abundant winter rains, infrequent snow, dry summers, and mild temperatures year-round (usually above freezing in winter, so that snow seldom remains for more than a few days). East of the Cascade crest, the region's climate is much more continental, with rainfall and cloudiness less common and sunshine and dry conditions more common, year-round. The annual and daily ranges of temperature east of the Cascades are considerably greater than those in the west. Winters are colder, snow is more common at low elevations, summer days are hotter, and summer nights are typically cooler. A greater fraction of precipitation falls in the warm half of the year, especially in May and June.

Locally important features of PNW climate are also directly influenced by gaps and low-elevation passes in the Cascade Mountains. In winter, for example, very strong easterly winds at subfreezing temperatures often flow down the Columbia River Gorge, which defines part of the border between Washington and Oregon states. These "coho" winds expose the Columbia Gorge (including population centers around The Dalles, Hood River and Portland, Oregon) to some of the harshest aspects of the continental climate of the interior PNW, such as ice storms and low-elevation snow storms. Similar continental weather reaches locations on the west-side of the Cascades near the U.S./Canada border, where the Fraser River valley connects the interior of British Columbia with Puget Sound, giving the city of Bellingham, Washington, significantly more freezing and snowy winter weather than other locations in the Puget Lowlands. Somewhat weaker though still important connections between the east-side continental climate and west side drainage basins are provided by numerous low-elevation passes in the Cascade range. Together, the fine scale structure in terrain and climate produce an important fine-scale structure in the biota of the west slopes of the Cascade mountain range.


Subsections

1.2 Patterns of Climate Variability in the Twentieth Century

1.2.1 PNW surface climate variability
For the PNW, there is strong regional coherence in both precipitation and temperature variability, and variations in time are more important than variations in space. This does not apply to individual storms, or weather variations over the course of a few days or weeks, but only to variations of a month or more.

The conclusion just stated is based on an analysis of US climate division data using a computational tool known as "principal component analysis" or PCA. Principal Component Analysis of cool-season (October-March averaged) data from all of the region's climate divisions reveals that the dominant pattern (the first Principal Component) of year-to-year temperature variations explains 84% of the variance and is regionally coherent. For precipitation, the dominant cool-season pattern explains 78% of the variance and is also regionally coherent. That is, there is a strong tendency for the regional temperatures and precipitation to vary as a whole. Although these results are based on station data that have been aggregated over PNW sub-regions, they are consistent with the results of PCA applied to a collection of distributed station data in the PNW. Thus, variations in space (across the region) are less important than variations in time (of the whole region). We take advantage of the regionally coherent variations of PNW climate in our subsequent descriptions, ignoring the subtler spatial variations within the region.

Different types of variables can be considered together using PCA to identify common patterns of variation in space and time. For our analysis of winter climate, the list of variables is expanded to include snowpack, water year (October-to-September) streamflow, and coastal ocean sea surface temperatures (SST), in addition to air temperature and precipitation. Perhaps not surprisingly, the resulting patterns of coherence are dominated by an amplified hydrologic response to fluctuations between warm-dry and cool-wet winters. The leading PC from this analysis (PC1) explains about 46% of the total variance in the input data for the 1946-1995 period of record. The second PC (PC2) explains about 25% of the total variance, capturing winter climate variations that are either anomalously warm-wet or cool-dry. Although PNW winter temperature and precipitation are essentially uncorrelated, cool-wet winters produce a large snowpack in the mountains and abundant water year streamflow, while warm-dry winters produce a relatively small snowpack and below average water year streamflow. These results suggest that the regional surface water supply (as measured by snowpack and streamflow) is sensitive to both precipitation and temperature: hydrologic response is greatest for winters that are both cool and wet, or both warm and dry. In contrast, the hydrologic response is muted during winters that are both cool and dry, or both warm and wet, relationships that are borne out by the weak correlations between PC2 and PNW streamflow and snowpack records.

Principal Component analysis of warm-season (April-September) data consisting of temperature, precipitation, and coastal SST also finds most of the variance concentrated in regionally coherent patterns. The first PC for warm-season data finds that the see-saw tendency for summers to be either warm-dry or cool-wet, across the entire PNW region, explains the largest fraction of the input data variance (about 41%). The second PC captures the somewhat less frequently observed fluctuations between cool-dry and warm-wet summers for the region (explaining 28% of the variance in the 1946-95 period of record).

We get a sense for the temporal rhythms in PNW climate by using indices such as PC1 and PC2 to track year-to-year climate conditions. Though there are clearly many climate-related measures of interest, we limit our discussion to PC1 and PC2 - the multivariate indices produced by this PCA - because they track several important climate variables at the same time.

As described above, winter PC1 is a good indicator for fluctuations between warm-dry low snowpack and low streamflow years, versus cool-wet high snowpack and high streamflow years. The time history for winter PC1 and a much simpler, albeit similar, 3-component Pacific Northwest climate index (developed by Curtis Ebbesmeyer) are shown in the top panel of Figure 4. PC_timeseries (insert it here). Ebbesmeyer's index was designed to highlight fluctuations between cool-wet high snowpack winters and warm-dry low snowpack winters by averaging normalized indices for winter precipitation at Cedar Lake (located in the western foothills of the central Washington Cascades), plus the April 15 snow depth at Paradise Ranger Station (on Washington's Mount Rainier), minus the annual air temperature anomaly in the San Juan Islands at Olga, Washington. By construction, Ebbesmeyer's index mimics PC1 because it has large positive values for cool-wet winters, and large negative values for warm-dry winters, in western Washington state. Both Ebbesmeyer's index and our winter PC1 have large positive amplitudes (indicating that cool-wet winters prevailed) from about 1946 to the late 1950's, then again from 1970 to 1976. Large negative amplitudes indicating relatively warm-dry winters, low snowpack and below average water year streamflow were common from 1940 to 1945, and again from 1977 to 1994.

1.2.2 Atmospheric circulation patterns that influence PNW surface climate variability
To identify the large scale atmospheric patterns that cause the regional climate variations described above, we construct circulation maps by simply using our time series (PC1 and PC2) to categorize years in the historic record, and then making "composites" of past climate data within the categories of interest. Composite cool-wet winter circulation maps, for example, are developed by averaging gridded sea level pressure (SLP) and 500 hPa height data for all available years having normalized winter PC1 values greater than +0.5 (left column of Figure 5). These fields are shown as anomalies from the 1948-97 long-term mean.

The composite cool-wet 500 hPa height anomalies (top left panel) bear a strong resemblance to a well-known cool-season circulation pattern of variability, the Pacific/North America (PNA) circulation pattern. The hemispheric scale wave-like pattern of tropospheric height anomalies for the cool-wet PNW winter composite - below average 500 hPa heights over northwestern North America and the subtropical Pacific from Hawaii westward, and above average heights over the Aleutians - is nearly identical to that of the PNA pattern. The SLP anomalies are essentially co-located with those at 500 hPa, with anomalously high SLPs over the North Pacific centered on the Aleutians - resulting in a weaker than average Aleutian Low pressure cell - and low SLP in western North America and in the subtropics west of Hawaii. For the PNW, these patterns support enhanced counterclockwise (cyclonic) flow and an active storm track following a more westerly/northwesterly track than that of the long-term climatology.

This weak Aleutian Low (negative PNA) circulation pattern typically brings relatively cold storms to the PNW that produce heavy precipitation and plentiful snowfall on the west slopes of the mountains. The hemispheric nature of this circulation pattern tells us that, when the Aleutian Low is weak, cool-wet weather in the PNW tends to coincide with dry weather in Hawaii, warm-dry weather in the southern U.S., and cool weather throughout western Canada and Alaska.

Composite maps for the warm-dry extremes of winter-PC1 are nearly perfect negative images of those for the cool-wet extremes, and are therefore not shown. The circulation patterns in the warm-dry case depict an intense Aleutian Low and ridging high pressure over the PNW region, a combination that tends to block precipitation-bearing storms while carrying relatively warm air into the region. Again, the hemispheric nature of this circulation pattern tells us that when it is bringing warm-dry weather to the PNW, it is also favoring wet weather in Hawaii, cool-wet weather in the southern U.S., and warm temperatures throughout western Canada and Alaska.

The October-March circulation patterns that consistently bring cool-dry PNW winters are shown in the right column of Figure 5. In this case, the presence of a strong ridge of anomalously high pressure over the Gulf of Alaska is featured in both the 500 hPa and SLP composite fields, as is an area of anomalously low heights (and SLPs) over the Bering Sea and the southwest U.S. During winter, this circulation pattern supports an anomalously cool and dry offshore flow, i.e., a greater propensity for continental air masses, over the entire PNW region. Note that the amplitudes of the height and SLP anomalies in this composite are about half those seen in the composite for PC1 and the circulation anomalies are more confined to the North Pacific sector than those associated with PC1.

The composite 500 hPa and SLP fields for the warm-wet years of PC2 (not shown) are essentially negative images of those for the cool-dry composite (Figure 5). Again, the flow patterns associated with the warm-wet circulation fields are consistent with the PNW climate anomalies of interest: enhanced southwesterly flow and a southwesterly storm track over the PNW region brings relatively warm-and-wet weather.

Subsections

1.3 The El Niño/Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO): Major Influences on PNW Climate

As implied by the composite circulation patterns, PNW climate variability is strongly influenced by climate events in and above the Pacific Ocean. Further, the dominant pattern of the region's hydrologic variability is closely related to the Pacific North America (PNA) circulation pattern, well known by climate scientists because of its dominant role in seasonal to inter-annual North Pacific climate variations.

1.3.1 The ENSO and PDO Phenomena
Both ENSO and PDO are patterns of Pacific climate variability that include changes in sea and air temperatures, winds, and precipitation. ENSO is Earth's dominant source of year-to-year climate variations. This phenomenon is understood to be a natural part of climate that spontaneously arises from interactions between tropical trade winds and ocean surface temperatures and currents near the equator in the Pacific.

On average, a cold tongue of upper ocean waters extends from the coast of South America into the central equatorial Pacific, while a warm pool of upper ocean waters exists in the western tropical Pacific. Relatively steady easterly trade winds blow from the cool eastern tropics to the warm western tropics, and the surface winds pick up warmth and moisture as they travel from east to west. Consequently, the warmed and moistened low-level winds become buoyant, ultimately fueling the frequent development of deep clouds that drop heavy precipitation over the warm pool sector of the western tropical Pacific.

El Niño, or the warm phase of ENSO, refers to periods in which the eastern and central equatorial Pacific Ocean is warmer than average. During El Niño, the easterly trade winds are weak, and tropical Pacific rainfall shifts eastward and equatorward from its usual locations. The shifting rainfall patterns that come with El Niño often bring drought to eastern Australia and much of southeast Asia while at the same time bringing drenching rains and flooding to typically dry places like islands in the central Pacific and the coastal deserts in northern Peru.

La Niña, or the cool phase of ENSO, tends to produce a stronger version of the average tropical circulation: intense easterly trade winds, unusually cool ocean temperatures in the eastern and central equatorial Pacific, abundant rainfall in eastern Australia and southeast Asia, and very dry weather over the typically dry places in the tropical Pacific.

Although El Niño and La Niña (ENSO extremes) are tropical phenomena, they often influence climate around the world. Simply put, the massive changes in tropical rainfall that are a part of El Niño and La Niña cause major changes in the atmosphere's heat engine. The altered tropical heating radiates disturbances to higher latitudes in wave-like patterns, and these can influence winds and storm tracks thousands of kilometers away. It is these wave-like disturbances high in the troposphere that allow ENSO variations to influence North America's climate.

The ENSO influence on North Pacific and North American climate is especially strong in the months from October through March, largely because the seasonal characteristics of northern hemisphere winds allow the tropically generated waves to propagate through the upper troposphere to higher latitudes. In contrast, summertime wind patterns in the Northern Hemisphere are such that they generally absorb the tropically generated wave energy, effectively trapping the ENSO-related disturbances in the tropics.

Like ENSO, PDO is a hemispheric-scale seesaw pattern in Pacific climate, but with several important differences. First, PDO appears to have its strongest signature in the North Pacific, instead of the tropical Pacific. Figure 6 shows the sea surface temperature (SST) anomalies that are associated with the warm phases of PDO and ENSO. The spatial patterns are very similar: both favor anomalously warm sea surface temperatures near the equator and along the coast of North America, and anomalously cool sea surface temperatures in the central North Pacific. (Because these patterns were derived from a linear analysis, the cool phases for PDO and ENSO, which are not shown, simply have the opposite patterns of SST anomalies: cool along the equator and the coast of North America, and warm in the central north Pacific.)

Tracking the ENSO and PDO climate patterns is often done by comparing observed climate conditions with the signature patterns described above, usually on a monthly averaged basis. For ENSO, a commonly used index is based on SSTs in the Niño3.4 region (5ºN-5ºS, 120ºW-170ºW), and we use it here. The PDO index is based on comparisons between monthly observations and the leading pattern of North Pacific SST variability (as determined from a Principal Component analysis, see Zhang et al. 1997; Mantua et al. 1997). Monthly values for the Niño3.4 and PDO indices are shown in Figure 7.

Differences in the temporal rhythms of ENSO and PDO are obvious: the lifetime of a typical ENSO event ranges from 6 to 18 months, and complete ENSO cycles typically have a 2 to 7 year period; by contrast, major PDO events in the twentieth century have stayed in one phase or the other for 20 to 30 years at a time, yielding a 50 to 70 year period for a complete PDO cycle.

A final key difference between ENSO and PDO lies in the state of current scientific understanding of the two phenomena. Scientists are reasonably agreed that ENSO exists because of strong air-sea interactions that take place in the tropical Pacific. These interactions give rise to a deterministic set of climate events that are inherently predictable at lead times of at least one to a few seasons. While ENSO has been extensively studied and is now routinely predicted at more than a dozen centers around the world, the causes for, and the potential ability to predict, PDO variations are not currently known. Part of the difficulty in understanding PDO results from the fact that its period is so long, compared to the period of good instrumental records in the North Pacific (since about 1900), that only two complete PDO cycles have been observed. The PDO was in its cool phase from about 1890 to 1925 and from 1945 to 1977. It was in its warm phase from 1925 to 1945 and from 1977 to at least the mid-to-late 1990's. Pacific climate changes in the late 1990's have, in many respects, suggested another reversal in the PDO (from "warm" to "cool" phase conditions). However, a lack of PDO understanding makes it impossible to determine true "PDO reversals" soon after they occur.

1.3.2 Impacts of ENSO and PDO on PNW Climate
Because ENSO and PDO influence the atmospheric circulation over the North Pacific and North America, they are important factors for Northwest climate. A number of studies have shown that cold phases of both PDO and ENSO (La Niña) favor a weak cool season (October-March) Aleutian Low. Recall that a weak Aleutian Low is the circulation pattern linked to the cool-wet extremes in PNW climate, as well as the Pacific/North America surface climate anomalies noted to coincide with cool-wet winters in the PNW. Likewise, warm phases of ENSO (El Niño) and the PDO favor an especially strong cool season Aleutian Low, the circulation pattern linked to the warm-dry extremes in PNW cool season climate. Although there are notable differences between the circulation patterns favored by ENSO and PDO, the differences are relatively unimportant for PNW climate.

One way to quantify the ENSO and PDO connections to PNW climate is to compare measures for these climate fluctuations with measures for the region's climate and hydrologic variability. We do so by determining the correlations between indices for the hemispheric scale patterns with the PCs from our analysis of PNW climate data discussed earlier. Recall that, in both halves of the year, the first two PCs together explain about 70% of the variance in the regional climate data matrices. The correlation between winter PC1 (the cool-wet versus warm-dry climate index) and the PDO is -0.64; the correlation between PC1 and Niño3.4 is also relatively large (-0.50) (Table 1.2).

Table 1.2 Results of principal components analysis of PNW climate for October-March (cool season) and April-September (warm season). The top row lists the percent variance explained by each pattern of coherence; the lower two rows show the correlation of the principal component time series with each season's PDO and ENSO index. Statistically significant correlations are shown in bold face.


 

October-March

April-September

 

PC 1

PC 2

PC 1

PC 2

%variance

41%

27%

41%

28%

PDO

-0.64

-.04

-0.45

-0.40

ENSO

-0.50

.03

-0.26

-0.32

The negative sign of the correlations indicates that the warm phases of ENSO and PDO tend to coincide with the warm-dry PNW winters, while the cool phases of ENSO and PDO show a strong association with cool-wet PNW winters. The second winter pattern (PC2) is essentially uncorrelated with the Niño3.4 and PDO indices, meaning there is no linear association between this PNW climate pattern and ENSO or PDO.

Another view of the ENSO and PDO influences on PNW climate is provided by warm and cool phase composites of PNW climate records, such as those shown in Figure 8. Here, we have selected monthly averaged surface temperature and precipitation for the PNW (Oregon, Washington, and Idaho). The largest differences in El Niño versus La Niña year composites occur in the cool season months. The composite El Niño surface temperature is ~0.4 to 0.7 ºC higher, on average, than the composite La Niña surface temperature in December through June. From October to March, composite El Niño year precipitation is on average about 1 cm per month less than in the La Niña year composite. Overall, the El Niño composite October-March precipitation is 14% less than that in the La Niña composite.

PNW temperature and precipitation composites based on cool and warm extremes of the PDO have their largest differences in the fall, winter and spring seasons. The warm PDO composite has October-to-May temperatures that are on average ~0.5 ºC higher than those in the cool PDO composite, with a range of +1.3 ºC in March-April to -0.25 ºC in November-December (Figure 9). October-January precipitation is ~1.2 cm per month lower in the warm PDO composite than in the cool PDO composites. Overall, the water year precipitation in the warm PDO composite is ~10% less than that in the cool PDO composite.

As might be expected, the combined influences of cooler-wetter climate in La Niña and/or cool phase PDO years favors higher snowpack and water year streamflows than during El Niño and/or warm phase PDO years.

 

 

1.4 Implications for Climate and Resource Predictions

Given our current understanding of PNW climate variability and its relationships with hemispheric climate patterns, what are the prospects for making skillful climate predictions for the PNW? As detailed above, there is ample evidence that hemispheric scale climate variations associated with the PDO and ENSO exert important influences on PNW climate, especially in the months October-March. Perhaps it should come as no surprise that the dominant pattern of PNW winter climate variability (PC1) captures much of the region's response to El Niño-Southern Oscillation (ENSO) and Pacific Decadal Oscillation (PDO) climate influences.

The results of our retrospective analyses suggest that an ability to monitor and predict the rhythms of ENSO and/or PDO could be exploited to predict swings between the cool-wet and warm-dry PNW pattern, purely on an empirical basis. On the other hand, the ability to predict aspects of PNW climate not captured by the region's dominant winter pattern may not be as promising, at least via empirically determined relationships. Because swings between warm-wet and cool-dry winter climate in the PNW have historically shown no temporal coherence with ENSO or the PDO, the ability to predict variations in this secondary pattern of PNW climate variability is unclear.

At the seasonal to inter-annual time scale, skill in monitoring and predicting variations in the ENSO cycle has been demonstrated for nearly a decade. So-called "two-tiered" climate forecasts are now routinely used to extend ENSO predictions to climate predictions for other regions, including the PNW. Once an ENSO forecast is made, relatively simple ENSO-related predictions can be generated from empirical relationships based on historic climate data; more sophisticated approaches first require a prediction for tropical SSTs that are later used as boundary conditions in climate model simulations. In the early stages of the 1997-98 El Niño - as early as June 1997 - both approaches were employed to make remarkably accurate forecasts for the 1998 winter and spring climate of North America.

As a general principle, all climate forecasts are probabilistic. For example, a typical El Niño-related climate forecast for the PNW might be presented as follows:

"Based on expectations for continued El Niño conditions in the tropical Pacific, we also expect increased likelihoods for above average winter and spring temperatures with below average precipitation, with small but non-zero odds for the opposite conditions (i.e., below average likelihoods for below average winter and spring temperatures and above average precipitation)."

The fact that correlations between ENSO indices (like the Niño3.4 index) and PNW climate indices are far from perfect brings up an important point, one often overlooked when anticipating the impacts of a predicted El Niño or La Niña event: The PNW climate anomalies that accompany warm or cool ENSO events are not always the same - in fact, there are many examples of winters that defied the "typical" pattern.

Some of the event-to-event differences in the PNW climate response appear to depend on the strength of the ENSO event in question. Although the sample size for extreme El Niño events is small, there is evidence supporting the notion that PNW winter climate tends to warmer temperatures but near-normal precipitation conditions during these rare cases. El Niño events in 1982-83 and 1997-98 were the most intense observed in this century, as measured by tropical ocean temperature, wind, and rainfall anomalies. Instead of simply causing an intensification in the typical El Niño-favored warm-dry PNW climate pattern, these two extreme El Niño events coincided with well above average winter and spring temperatures but near or above average precipitation across the region (Table 1.3). The net result was that snowpack and streamflow anomalies following the winters of 1982-83 and 1997-98 were smaller than those typical of an average intensity El Niño event.

Table 1.3 Comparison between PNW temperature and precipitation anomalies during a composite of El Niño years in the 1946-99 period of record (excluding the extreme 1982-83 and 1997-98 events) and those observed during 1982-83 and 1997-98. Data comparisons made for October-March averages based on US Climate Division Data for Idaho, Oregon, and Washington.


 

Surface temperature anomaly

Precipitation anomaly (percent of normal)

Composite El Niño year

+0.5 C

93%

1982-83

+0.7 C

125%

1997-98

+1.2 C

101%


Differences in PNW climate anomalies during the extreme and more typical El Niño episodes can be traced to the character of the accompanying atmospheric circulation anomalies. Composite October-March circulation anomalies for El Niño events between 1946 and 1996 (excluding the 1982-83 event) feature a deepened Aleutian Low, enhanced ridging over northwestern North America, and troughing over the southeast U.S.. This pattern has a clear projection on the PNA pattern (and on the circulation that favors warm-dry periods for the PNW). In contrast, during October-March of 1982-83 and 1997-98 the Aleutian Low was exceptionally deep and shifted southeastward from its climatological position (Figure 10), and greatly enhanced ridging was centered over Hudson Bay and the Great Lakes. These circulation anomalies would bring unusually warm moist air into the Pacific coast states and favor near-to-above average precipitation for the PNW.

A few months prior to the winter of 1998, several state-of-the-art climate prediction models did an excellent job in forecasting the extreme intensity and southeastward displacement of the Aleutian Low (and other important aspects of North American climate) that were missed by empirically-based climate forecasts that called for "typical" El Niño conditions. Climate forecasting with physically based models, rather than empirical relationships, offers great potential for improving forecasting skill because the models are not constrained by historical relationships. Present day forecasting centers are generally increasing their reliance on models while still making extensive use of empirical relationships.

As previously noted, there is currently little demonstrated skill in predicting PDO variations. This situation is directly related to the fact that the mechanisms giving rise to the PDO are not understood. Although some climate simulation models produce PDO-like oscillations, they often do so for different reasons. The mechanisms giving rise to PDO will determine whether skillful decades-long PDO climate predictions are possible. If PDO arises from air-sea interactions that require ten-year ocean adjustment times, then aspects of the phenomenon will (in theory) be predictable at lead times of up to 10 years.

Even in the absence of a theoretical understanding, PDO climate information improves season-to-season and year-to-year climate forecasts for North America because of its strong tendency for multi-season and multi-year persistence. Simply assuming persistence of observed PDO-related North Pacific SST anomalies in the fall in any given year provides some skill in predicting PDO-related winter climate anomalies in the PNW region. NOAA's Climate Prediction Center has exploited this facet of North American climate with their "Optimal Climate Normals" (OCN) statistical prediction tool. In the absence of El Niño or La Niña, assuming persistence in the observed PDO state provides much of the skill in seasonal climate forecasts for North America. However, this persistence based forecast will always fail to predict the relatively infrequent switches from one PDO phase to another.

Combining ENSO and PDO information offers a promising means of maximizing basin scale climate information for use in predicting PNW climate (and North American climate more generally). The separation between probability density functions (PDFs) and composites of PNW climate data, when segregated according to ENSO and PDO extremes simultaneously, is largest when the two large-scale climate patterns are in the same phase (either warm/warm or cold/cold). In contrast, when the ENSO and PDO are in opposing phases (either cool/warm or warm/cool), Cascade Mountain snowpack and other water cycle parameters tend to be near their climatological values (see the Water Resources Theme Page).

Tendencies for temperature and precipitation anomalies to sometimes covary in predictable ways offers a means for making skillful predictions for snowpack, streamflow, and other resources sensitive to the water cycle. Hamlet and Lettenmaier (1999, 2000) have developed a methodology for extending the lead time of water resources forecasts for the Columbia Basin by selectively resampling the historic climate record based on forecasts for ENSO and PDO. Thus, climate forecasts can provide a basis for making resource forecasts. The full value of climate forecasts can only be realized with the added value predictions that go beyond forecast products like probability outlooks for temperature or precipitation. The value of climate information increases when the expected physical climate anomalies are translated into expected impacts on tangible resources like water supply, hydropower, fish, and forests.

It is important to bear in mind that perfect predictions for the dominant hydrological climate pattern (winter PC1) would have captured about 45% of the total variance in PNW climate in the twentieth century. For this time period, variations in ENSO and PDO accounted for about 60% of the variance in winter PC1. Therefore, perfect predictions for ENSO and PDO would have accounted for about 27% of the total variance in our twentieth century October-March PNW climate data - even in this best-case scenario about 70% of the region's winter climate variance remains unexplained. There are indications that both improved climate modeling and climate diagnostics will allow for predictions that reduce the fraction of unexplained climate variance.

Finally, "noise" and the chaotic nature of earth's climate will always place strict limits on both the potential and realized skill in climate predictions. The exciting message from recent successes in climate prediction is that opportunity forecasts associated with major climate events - like the El Niño of 1997-98, and the La Niña event of 1998-2000 - can predictably alter the odds for future climate conditions in select regions. As far as seasonal to interannual climate variations are concerned, the PNW is situated in one of the more predictable parts of North America, with the added bonus of having an amplified response in the region's water cycle to the most predictable swings in Pacific/North America winter climate.


   
 

 


Back to top | CIG Home