Abstract

Atmospheric circulation anomalies and hydrologic processes associated with California wet and dry events were examined during Northern Hemisphere winter. The precipitation anomaly over the west coast of North America shows a north–south three-cell pattern. Heavy precipitation in California is accompanied by dry conditions over Washington, British Columbia, and along the southeastern coast of Alaska and reduced precipitation over the subtropical eastern Pacific. The inverse relationship between California and the Pacific Northwest is supported by the transport of moisture flux. During wet events, the southern branch of moisture flux transport strengthens and brings moisture from the North Pacific to California, hence enhanced rainfall. Strengthened moisture flux transport northward to the area north of Washington is consistent with suppressed rainfall in California.

The local precipitation anomaly pattern in the eastern tropical Pacific just north of the equator has a large influence on precipitation events in California. The enhanced precipitation generates strong rising motion. The associated sinking motion is located over California. Strong sinking motion and strong upper-level convergence favor dry conditions in California. Conversely, suppressed rainfall in the eastern Pacific is associated with above-normal precipitation in California.

Precipitation in California is likely below normal during cold ENSO events. When convection in the central Pacific is enhanced, California has heavy precipitation if rainfall in the subtropical eastern Pacific is suppressed. In addition to ENSO, precipitation in California is also modulated by the tropical intraseasonal oscillation. Wet (dry) events are favored during the phase of the oscillation associated with enhanced convection near 150°E (120°E) in the tropical Pacific.

1. Introduction

California normally receives the bulk of its precipitation during Northern Hemisphere (NH) winter from October to April when the storm track across the North Pacific is active. Both orography and land–sea contrast contribute to regional characteristics of precipitation, but persistent wet and dry episodes are usually associated with persistent circulation anomalies influenced by both local and remote boundary forcing. Ropelewski and Halpert (1986) examined the relationship between El Niño–Southern Oscillation (ENSO) events and precipitation in the United States, but they did not find a consistent relationship between California precipitation and ENSO. Schonher and Nicholson (1989) found abnormally high precipitation in California during most of the warm ENSO years. Since not all warm events produce heavy precipitation, they further classified ENSO years according to the profile of sea surface temperature anomalies along the equator and established alinkage between precipitation in California and type 1 ENSO events as defined by Fu et al. (1986). These ENSO events have strong SST warming in the central and eastern Pacific and close to normal conditions in the western Pacific. Their results were confirmed by Cayan and Webb (1992) and Kahya and Dracup (1994), who related streamflows in the Pacific Southwest to type 1 ENSO events. However, there is a weakness in the relationship between precipitation in California and ENSO recognized by Cayan and Peterson (1989). Below normal precipitation usually was observed during the cold ENSO winters, but heavy precipitation also occurred during many non-ENSO winters, such as 1978, 1993, and the most recent floods of 1997.

To understand the characteristics of California precipitation, it is important to link the wet/dry conditions to the atmospheric circulation. Cayan and Redmond (1994) related the atmospheric teleconnection patterns (Wallace and Gutzler 1981) to precipitation in the western United States. Several teleconnection patterns such as the western Pacific oscillation, the Pacific–North American, the eastern Pacific, and the tropical–Northern Hemisphere patterns are linked to precipitation in the western United States. These patterns are also linked to ENSO (Livezey and Mo 1987). In addition to the teleconnection patterns, Cayan and Redmond (1994) alsorelated the westerlies along the west coast of North America to precipitation. When the westerlies extend to the west coast of the United States, precipitation is generally heavy, and when the westerlies are confined to the central Pacific, precipitation tends to be light.

Many previous studies have used monthly mean data to investigate California precipitation. However, wet or dry events are episodic and most events last less than one month and events often cross monthly boundaries. In this paper, we begin with wet and dry events selected from daily precipitation in California (Higgins et al. 1996a) and go on to examine hydrologic conditions and atmospheric circulation anomalies common to all wet (dry) events. We focus on southern California where rainfall is known to be influenced by tropical forcing (Cayan and Redmond 1994). We then classify wet and dry events according to the convection patterns in the tropical Pacific and present evidence that in addition to ENSO, wet and dry episodes are modulated by the tropical intraseasonal oscillation (IO).

In addition to the daily observed precipitation dataset, the dataset used in this study is the Climate Data Assimilation System/reanalysis from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) (Kalnay et al. 1996). Investigations of large-scale aspects of the hydrologic cycle revealed by reanalysis (Mo and Higgins 1996; Higgins et al. 1996b; Trenberth and Guillmot 1995) indicate, overall, that the analysis products are realistic especially in the data rich Northern Hemisphere. The reanalysis forecast products have biases, particularly in the 6-h precipitation forecasts. These biases, which are largely systematic, can be removed. Together with other independent data sources such as observed precipitation and outgoing longwave radiation (OLR), the reanalysis data provide useful information on relationships between precipitation and circulation features.

The datasets used in this study are described in section 2. The hydrologic cycle associated with California precipitation and the classification of wet and dry events are discussed in section 3. Large-scale atmospheric anomalies and hydrologic conditions common to wet (dry) events are examined in section 4. Wet (dry) events are classified into different categories based on tropical convection in section 5. We present evidence that precipitation in California is often modulated by the tropical IO in section 6. Conclusions are given in section 7.

2. Data

The data used in this study are daily mean global gridded analyses and forecasts from the NCEP–NCARreanalysis for the period from 1973 to 1995. The data are on a 2.5° lat × 2.5° long grid and 28 levels in the vertical. Precipitation (P) and evaporation (E) are taken from the 0–6-h accumulation from the forecast cycle. Moisture fluxes were computed directly from sigma level data (Mo and Higgins 1996). Daily averages of the NOAA satellite outgoing longwave radiation (OLR) field are used as a proxy for tropical convection (Liebmann and Smith 1996). The OLR data cover the period from 16 June 1974 to 31 December 1995 with 10 months (1 March–31 December 1978) missing. The seasonal cycle at each grid point is defined as the grand mean plus the annual and semiannual cycles. Anomalies are defined as departures of daily values from the seasonal cycle. Low-pass-filtered fields (periods greater than 10 days) are obtained by applying the low-pass filter of Blackmon (1976). Synoptic-scale eddies (periods between 2.5 and 6 days) are obtained by passing anomalies through a bandpass filter (Blackmon 1976). To obtain the IO signal, we filtered anomalies using the minimum bias window developed by Papoulis (1973) to retain periods in the range of 10–90 days.

Over the United States, daily observed precipitation derived from gridded hourly station data was used to select events and for verification. The data are on a 2°lat × 2.5° long grid covering the period from 1963 to 1995 (Higgins et al. 1996a). To examine precipitation anomalies in the eastern Pacific, we use values derived from the Geostationary Operational Environmental Satellite (GOES) precipitation index (GPI) pentad data from 1986 to 1995. The GPI technique is used to produce estimates of tropical rainfall for the Global Precipitation Climatological Project (Janowiak 1992). Since the GPI algorithm is generally valid in the Tropics, GPI values are used between 38°S–38°N. The precipitation anomalies are defined as departures from the mean daily (pentad) climatology.

3. Moisture transport associated with California precipitation

a. Winter climatology

The winter (January–March) mean precipitation averaged from 1979 to 1995 from the monthly mean global merged precipitation analysis of Xie and Arkin (1996) is given in Fig. 1d. Over the United States, values are comparable to the climatology of Higgins et al. (1996a). The main precipitation centers in the United States during winter are located in the northwest and the southeast. The maximum centered near Washington state is more prominent during early winter. Precipitation starts to spread southward in January, and California receives much of its precipitation from January to March. The southeast center persists through the winter months, and precipitation in the North Pacific is located at the exit region of the subtropical jet. In the Tropics, the ITCZ is centered at 5°N.

Fig. 1.

Mean winter (January–March) (a) precipitation from the NCEP–NCAR reanalysis. Contour interval 2 mm day−1. (b) Same as (a) but for vertically integrated moisture flux (arrow) and moisture divergence D(Q) (contoured). Negative values are shaded. Unit for moisture flux: 10 kg (m-s)−1. (c) Same as (b) but for PE. Positive values are shaded. (d) Observed precipitation from Xie and Arkin (1996). Contour interval 2 mm day−1.

Fig. 1.

Mean winter (January–March) (a) precipitation from the NCEP–NCAR reanalysis. Contour interval 2 mm day−1. (b) Same as (a) but for vertically integrated moisture flux (arrow) and moisture divergence D(Q) (contoured). Negative values are shaded. Unit for moisture flux: 10 kg (m-s)−1. (c) Same as (b) but for PE. Positive values are shaded. (d) Observed precipitation from Xie and Arkin (1996). Contour interval 2 mm day−1.

Precipitation from the reanalysis is known to havebiases especially in the eastern Pacific and over the landmasses (Mo and Higgins 1996; Trenberth and Guillemot 1995). However, the model is able to capture the observed precipitation along the west coast of North America, in the North Pacific, and over the ITCZ (Fig. 2a). It misses the maximum in the southeast and there is also too much rainfall over central America. The evaporation rate over the west is less than 1 mmday−1, so precipitation is mostly due to the vertically integrated moisture flux convergence D(Q). In the southeast, both evaporation and moisture flux convergence contribute. In the Pacific and over the western United States, PE (Fig. 2c) is well balanced by −D(Q). In the southeast, D(Q) shows a displaced maximum near Virginia. During winter, there are two distinct pipelines of moisture transport (Fig. 2b) that bring moisture to the west coast of the United States from the Pacific Ocean. The northern branch, which brings moisture to the Pacific Northwest dominates during early winter. During late winter (February and March), moisture flux is largely on the southern path and brings moisture to California and northern Mexico. The precipitation in the southeast originates from northward moisture transport over the Gulf of Mexico and also from the Pacific. The moisture flow turns northeastward and exits across the east coast of the United States. There is another center of moisture flux convergence over the ITCZ, but over the oceans the contribution of E is important.

Fig. 2.

(a) Precipitation difference between wet and dry events from the NCEP–NCAR reanalysis 6-h forecasts. Contour interval 2 mm day−1. Zero contours are omitted. Negative values are shaded. (b) Same as (a) but for vertically integrated moisture flux divergence and moisture flux difference between wet and dry events. Contour interval 2 mm day−1. Positive values are shaded. Unit for moisture flux 10 kg (m s−1)−1,. (c) Same as (a) but for PE. (d) Same as (a) but for observed precipitation from Higgins et al. (1996a,b).

Fig. 2.

(a) Precipitation difference between wet and dry events from the NCEP–NCAR reanalysis 6-h forecasts. Contour interval 2 mm day−1. Zero contours are omitted. Negative values are shaded. (b) Same as (a) but for vertically integrated moisture flux divergence and moisture flux difference between wet and dry events. Contour interval 2 mm day−1. Positive values are shaded. Unit for moisture flux 10 kg (m s−1)−1,. (c) Same as (a) but for PE. (d) Same as (a) but for observed precipitation from Higgins et al. (1996a,b).

b. Wet and dry events

We selected wet and dry events using gridded daily station data (Higgins et al. 1996a). Before selecting wet and dry events, we determined the precipitation distribution using the procedure outlined by Mo et al. (1997).Because of the skewness of the precipitation distribution, Richman and Lamb (1985) suggested using the square root or log10 transformation in analyzing rainfall. The square root transformation (Lanzante and Harnack 1982) was used here. The distribution of 5-day mean precipitation for southern California from 32° to 37°N and west of 118°W during the extended winter (1 November–31 March) was obtained by pooling all grid points in that area. There are 32 winters and, by pooling points together, the sample size is large enough to obtain stable statistics. To select events, we formed a time series of 5-day running mean precipitation summed over the above area for each winter from 1963/64 to 1994/1995. We then computed precipitation percentiles using the distribution function determined above. We are interested in persistent wet and dry episodes because they are associated with persistent circulation features influenced by the Tropics. We selected extreme precipitation events using the threshold crossing procedure outlined by Dole and Gordon (1983) with one modification. The magnitude criterion was replaced by a percentile requirement. Wet (dry) events were chosen when the square root of 5-day running mean precipitation is above the 80th percentile (below the 20th percentile) for at least 10 consecutive days. The duration of an event is defined as the period from the first crossing of the threshold (onset) to the next crossing of the threshold (demise). Over the 23-yr period (1973–95), there were 40 wet events with an average duration of 15 days. There were 32 dry events with an average duration of 19.4 days. We then obtained composite fields for wet (dry) events by averaging over the duration of each event and then averaging over all wet (dry) events. Statistical significance of each composite was assessed by assuming a normal distribution with the decorrelation time assumed to be 10 days. For precipitation, a square root transformation was applied before assessing the statistical significant of the composites. The circulation features in the Pacific North American area associated with wet events are similar to those associated with dry events with a sign reversal, so we present the composite difference between wet and dry events.

The composite difference of precipitation anomalies between wet and dry periods (Fig. 2d) from the observed station data (Higgins et al. 1996a) shows a seesaw pattern between the northern and southern regions along the western states. The positive anomaly center is concentrated over California as expected and rainfall extends from California to the Great Basin. The wetness in California is accompanied by negative anomalies over Washington, British Columbia, and along the southeastern coast of Alaska. All centers are statistically significant at the 95% level. There are only two stations along the coast of Alaska, but the seesaw pattern is reproduced by precipitation anomalies from reanalysis (Fig. 2a) and 200-hPa divergence anomalies (Fig. 3c).

Fig. 3.

(a) OLRA difference between wet and dry events. Contour interval 4 W m−2. Zero contours are omitted. Areas where values are statistically significant at the 95% level are shaded. (b) Same as (a) but for vertical velocity anomaly difference at 500 hPa. Contour interval 0.02 Pa s−1. (c) Same as (a) but for 200-hPa divergence anomaly and divergent wind (arrow). Unit for divergence 10−6 s−1. (d) Same as (a) but from the GPI pentad data. Contour interval 1 mm day−1. (e) Same as (a) but for sea level pressure difference. Contour interval 2 hPa. Contours −0.5 and 0.5 hPa are added. (f) Same as (e) but for 200-hPa kinetic energy anomaly composite for bandpass filtered (2.5–6 days). Contour interval 6 (m s−1)−2.

Fig. 3.

(a) OLRA difference between wet and dry events. Contour interval 4 W m−2. Zero contours are omitted. Areas where values are statistically significant at the 95% level are shaded. (b) Same as (a) but for vertical velocity anomaly difference at 500 hPa. Contour interval 0.02 Pa s−1. (c) Same as (a) but for 200-hPa divergence anomaly and divergent wind (arrow). Unit for divergence 10−6 s−1. (d) Same as (a) but from the GPI pentad data. Contour interval 1 mm day−1. (e) Same as (a) but for sea level pressure difference. Contour interval 2 hPa. Contours −0.5 and 0.5 hPa are added. (f) Same as (e) but for 200-hPa kinetic energy anomaly composite for bandpass filtered (2.5–6 days). Contour interval 6 (m s−1)−2.

The reanalysis captures the seesaw between the northern and southern regions in the west (Fig. 2a). In addition, it shows another seesaw between precipitation in California and dryness over the subtropical eastern Pacific centered at 12°N. The seesaw between California rainfall and the Pacific Northwest is supported by moisture flux transport (Fig. 2b). Contributions from evaporation difference are small so PE is dominated by precipitation. The PE difference (Fig. 2c) is balanced approximately by the vertically integrated moisture flux divergence difference (Fig. 2b), which shows a same pattern with imbalances of less than 1 mm day−1 at all three centers. For wet events, there is an enhanced moisture flux transport from the Pacific to California and Nevada (Fig. 2b). Enhanced moisture flux convergence is consistent with excessive precipitation in California. The situation reverses for dry events. The strengthening of moisture flux transport northward means less moisture being transported to the southwest, hence less rainfall.

4. Circulation anomalies and local tropical influences

The above three cell patterns of precipitation and moisture transport are supported by circulation anomalies. Composites of OLR anomalies (OLRA) have large variability among events in the Tropics, but locally, all events show a north–south three-cell pattern near the west Pacific coast of North America (Fig. 3a). The OLRA difference is consistent with 200-hPa divergent winds (Fig. 3c), where divergence (convergence) is located in the area of negative (positive) OLRA. Negative OLRA over California and positive anomalies over British Columbia and Washington are consistent with precipitation anomalies (Fig. 2a). The center of positive OLRA in the eastern Pacific near (12°N, 140°W) can be verified using the pentad GPI data. There were 12 wet and 13 dry events for the period from 1986 and 1995. The composite difference between those wet and dry events (Fig. 3d) indicates that wet events are associated with less rainfall in the subtropical eastern Pacific from 150°W to central America.

This seesaw between rainfall in California and the subtropical eastern Pacific just north of the ITCZ is supported by dynamics. Less precipitation in the eastern Pacific is associated with less rising motion as indicated by a reduction in vertical velocity difference of ∼0.02 Pa s−1 at 500 hPa (Fig. 3b). This suppresses the compensating subsidence located near California as indicated by the divergent wind (Fig. 3c). Less sinking motion in California creates a favorable condition for enhanced moisture flux convergence (Fig. 2b), hence more rainfall. The situation reverses for dry events. Enhanced rainfall in the area just north of the ITCZ generates strong rising motion. The associated strong sinking motion and suppressed moisture flux convergence located in California favor dry events. These common features suggest the importance of the local dynamical supportand that precipitation in California is influenced by tropical convection in the eastern Pacific.

The three-cell pattern also appears in the sea level pressure difference (Fig. 3e). It shows pressure drop (increase) in the area of enhanced moisture flux or low-level wind convergence (divergence). The pattern is also consistent with the anomalous jet stream reported by Cayan and Redmond (1994). They found that the westerlies extend to the west coast and merge with the North American jet during heavy California rainfall events. For dry events, the westerlies are confined to the central North Pacific, and along the west coast of North America, the jet stream shifts northward with a minimum located over California.

The locations of jet streams influence the locations of enhanced storm tracks represented by the bandpass-filtered eddy kinetic energy per mass at 200 hPa (Fig. 3f), which shows positive anomalies across the United States with a maximum located over California, which is consistent with enhanced precipitation. The southward shift of the storm tracks leaves Washington and British Columbia dry.

5. Further classification of wet and dry events

The OLRA among wet and dry events in the Tropics have large variability. The wet and dry events may be further classified according to the convection patterns in the central Pacific. After averaging OLRA over the duration of each event, we then classified wet and dry events according to the location of enhanced convection (the OLRA minimum) in the tropical Pacific. Wet events with enhanced convection east (west) of the date line are classified as category 1 (category 2) event; similarly for dry events. There are four wet events and six dry events that cannot be classified in this way and will not be discussed here.

For each category, we obtained composites of OLRA, 200-hPa zonal wind, and eddy streamfunction anomalies with zoanl means removed by averaging the anomalies over the mature phase (day 0 to day 10) of each event and then averaging over all events in the same category. The statistical significance for each composite is assessed by assuming a normal distribution. The decorrelation time is taken as 10 days. Areas where values are statistically significant at the 95% level are shaded. We discuss circulation anomalies associated with each category below.

a. Tropical convection located east of the date line—Category 1 events

1) Wet events

There are 10 events in this category with an average duration of 21 days. All 10 events occur during strong warm ENSO events such as 1976, 1982, 1992, and 1995. Warm events have been linked to California rainfall by Schonher and Nicholson (1989) and Cayan and Redmond (1994). The OLRA composite (Fig. 4a) shows a typical ENSO signal. The east–west OLRA dipole in the equatorial Pacific shows enhanced convection in the central and eastern tropical Pacific with a minimum located near 165°W, and suppressed convection in the western Pacific. Positive OLRA are also found in the subsidence regions of the local Hadley cell at 20°S and 15°N. In the Pacific near the west coast of the United States, a familiar three-cell pattern appears (Fig. 3a). The mean OLRA for each event were examined separately. The amplitudes of OLRA and exact locations of maxima and minima may vary, but they all show the same dipole pattern in the tropical Pacific and the three-cell pattern along the west coast of the United States.

Fig. 4.

(a) OLRA composite for category 1 wet events. Contour interval 8 W m−2. Zero contours are omitted. Areas where values are statistically significant at the 95% level are shaded. (b) Same as (a) but for category 1 dry events. Contour interval 4 W m−2.

Fig. 4.

(a) OLRA composite for category 1 wet events. Contour interval 8 W m−2. Zero contours are omitted. Areas where values are statistically significant at the 95% level are shaded. (b) Same as (a) but for category 1 dry events. Contour interval 4 W m−2.

2) Dry events

Enhanced convection in the central Pacific does not guarantee above-normal rainfall in California. There are also six dry events with similar convection pattern in the western and central tropical Pacific (Fig. 4b). The major difference between dry (Fig. 5b) and wet events (Fig. 5a) is in the eastern Pacific. The composite for dry events shows negative OLRA just north of the equator extending from the eastern Pacific to the Gulf of Mexico while the composite for wet events shows positive OLRA in the subtropical eastern Pacific. This seems to indicate that in addition to remote forcing, the local dynamics also plays an important role in determining rainfall in the west.

Fig. 5.

(a) OLRA composite for category 2 wet events. Contour interval 3 W m−2. Zero contours are omitted. Areas where values are statistically significant at the 95% level are shaded. (b) Same as (a) but for 200-hPa eddy streamfunction anomaly composite. Contour interval is 2 × 10+6 m2 s−1. Contours 1 × 10+6 m2 s−1 and −1 × 10+6 m2 s−1 are added. (c) Same as (a) but for 200-hPa zonal wind anomaly composite. Contour interval 2 m s−1.

Fig. 5.

(a) OLRA composite for category 2 wet events. Contour interval 3 W m−2. Zero contours are omitted. Areas where values are statistically significant at the 95% level are shaded. (b) Same as (a) but for 200-hPa eddy streamfunction anomaly composite. Contour interval is 2 × 10+6 m2 s−1. Contours 1 × 10+6 m2 s−1 and −1 × 10+6 m2 s−1 are added. (c) Same as (a) but for 200-hPa zonal wind anomaly composite. Contour interval 2 m s−1.

b. Tropical convection located west of the date line—Category 2 events

1) Wet events

There are 26 events with an average duration of 13.1 days. Events are almost evenly distributed through the years with no apparent connection to ENSO. The OLRA composite (Fig. 5a) shows the same north–south three-cell pattern along the west coast as the composite for the wet 1 events, but in the Tropics, patterns differ. Figure 5a shows negative OLRA extending from thewestern Pacific to the date line with a minimum located at 150°E and positive anomalies extending from the date line to the Gulf of Mexico. Positive OLRA are also found over the Indian Ocean. It resembles the composite of OLRA for the phase of the 30–60-day oscillation when positive OLRA are centered near 150°E (Lau and Chan 1985). In South America, an enhanced South Atlantic convergence zone is accompanied by dryness in the southern plains. Such a seesaw was also observed by Paegle and Mo (1997). Over the duration of some events, OLRA propagate eastward.

The composite of the 200-hPa eddy streamfunction anomalies with zonal means removed shows a dipole in the western Pacific with negative anomalies centered at 25°S, and weak positive anomalies centered at 20°N near the area of enhanced convection (Fig. 5b). The important feature is a wavetrain extending from the region of enhanced convection, passing through the North Pacific and the Gulf of Alaska to the western United States. In the western hemisphere, the composite resembles the Pacific–North American (PNA) teleconnection pattern which has been linked to heavy precipitation over the Southwest (Cayman and Redmond 1994). Anomalies in the eastern hemisphere are weak. The 200 hPa zonal wind anomaly composite (Fig. 6c) shows the strengthening of westerlies along the west coast of the United States. However, the wind anomalies are mostly local. The Pacific jet is at its climatological position.

Fig. 6.

Same as Fig. 5 but for category dry 2 events.

Fig. 6.

Same as Fig. 5 but for category dry 2 events.

2) Dry events

There are 20 events in this category and the average duration is 22 days. These events include six events during the cold ENSO winters (1974, 1994, and 1998), but events can also occur during non-ENSO winters when the tropical IO is strong. The OLRA composite (Fig. 6a) resembles the convection pattern during cold ENSO events and is similar to that for wet events during warm ENSO events (category 1) with a sign reversal. It shows enhanced convection located in the western Pacific near 120°–125°E, and suppressed convection in the central Pacific near 150°W, which is in quadrature with the OLRA composite for the wet 2 composite (Fig. 5a). These two patterns (Figs. 5a and 6a) represent two phases of the tropical IO (Lau and Chan 1985).

The 200-hPa eddy streamfunction anomaly composite (Fig. 6b) shows a dipole straddling the equator in the area of enhanced convection. Over North America, the pattern resembles the tropical Northern Hemisphere teleconnection pattern with positive anomalies extending from the west coast of the United States southeastward to Florida and positive anomalies in the Atlantic. The negative center over the central United States is only statistically significant at the 90% level. The retraction of the Pacific jet into the western Pacific (Fig. 6c) is consistent with suppressed convection in the central Pacific. The dipole located near the west coast indicates a northward shift of the jet stream. The North American jet is broader and extends farther into the North Atlantic.

6. Remote tropical influences

a. ENSO events

The above results indicate that wet (dry) events are favored during warm (cold) ENSO winters, which is consistent with findings of Schonher and Nicholson (1989) and Cayman and Redmond (1994) and many others. However, not all warm events produce heavy rainfall. The existence of dry 1 events indicates that enhanced convection in the central Pacific is not always linked to wet events. The control mechanism seems to be the precipitation anomalies in the subtropical eastern Pacific. When precipitation is suppressed in the subtropical eastern Pacific but is enhanced in the tropical central Pacific, then California receives above normal rainfall.

b. Linkages to the tropical intraseasonal oscillation (IO)

In addition to ENSO, California rainfall can be influenced by the tropical IO. The OLRA composites (Figs. 5a and 6a) for wet 2 and dry 2 events with enhanced convection west of the date line show east–west three-cell patterns in quadrature with each other in the tropical Pacific similar to those extended EOF patterns during the Madden-Julian oscillation in winter (Lau and Chan 1985). If the tropical connection to California rainfall is robust, we should be able to obtain similar OLRA and circulation anomalies (Figs. 5 and 6) based on composites keyed to OLRA in the Tropics. In this section, we use such composites to establish linkages between the tropical IO and California precipitation. We then focus on 10 days before onset when remote forcing may be important. We use lagged composites from day −10 to day 10 to present evidence that the tropical IO modulates California precipitation.

1) Composites based on tropical convection

Composites keyed to precipitation in California show that wet and dry category 2 events are associated with enhanced convection centered at 150°E (Fig. 5a), and 125°E (Fig. 6a), respectively. Because of missing data in OLRA, we used data only for the period from 1979 to 1995. We chose two 10° × 10° boxes centered at 5°S, 150°E and 5°S, 125°E as our base areas. We composite daily OLRA and 200-hPa eddy streamfunction anomalies when the OLRA averaged over the base areas are negative and exceed 1.0 standard deviation. Composites shown here are based on data from December to March. For the 16 winters, there are a total of about 430 (456) maps entering the composites. Among these maps, there are 123 (147) days (roughly 20%) in common with California wet (dry) events. There are many cases, positive (negative) precipitation anomalies over California do not persistent long enough or strong enough to be classified as a wet (dry) event. Since OLRA in midlatitudes may not represent rainfall, we also composite precipitation anomalies from gridded station data over the United States.

Figure 7 shows composites based on the box centered at 5°S, 150°E, which should be compared with composites keyed to California rainfall for wet 2 events (Fig. 5). When the base point is shifted from 10°S to 5°N and from 140° to 155°E, the pattern remains the same. All show negative OLRA over California. Both OLRA composites (Figs. 5a and 7a) show an east–west three-cell pattern in the tropical Pacific with enhanced convection centered at 150°E and cloud bands extending from the subtropical eastern Pacific (20°N) to the west coast of the United States. The negative OLRA correspond well with the positive precipitation anomalies over California (Fig. 7d). The precipitation composite also shows less rainfall in the Pacific Northwest. Most fluctuations are from the intraseasonal band. The composite of 10–90-day filtered OLRA anomalies (Fig. 7c) shows essentially the same pattern (Fig. 7a), but the magnitudes are about 20% weaker. The 200-hPa streamfunction anomaly composite (Fig. 7b) captures negative anomalies in California and positive anomalies near the Gulf of Alaska but centers are shifted. In the Tropics, both show a dipole straddling the equator near the area of convection west of the date line. Anomalies elsewhere are different from the composite keyed to California precipitation (Fig. 5b).

Fig. 7.

(a) Composite of OLRA based on the tropical OLRA averaged over a 10° × 10° box centered at 5°S, 150°E. Contour interval 5 W m−2. Zero contours are omitted. Contours −2 and 2 W m−2 are added. Areas where values are statistically significant at the 95% level are shaded. (b) Same as (a) but for 200-hPa eddy streamfunction anomaly composite. Contour interval 1 × 10+6 m2 s−1. (c) Same as (a) but for 10–90-day filtered OLRA. (d) Same as (a) but for precipitation anomaly composite over the United States. Contour interval 0.3 mm day−1. Positive values are shaded.

Fig. 7.

(a) Composite of OLRA based on the tropical OLRA averaged over a 10° × 10° box centered at 5°S, 150°E. Contour interval 5 W m−2. Zero contours are omitted. Contours −2 and 2 W m−2 are added. Areas where values are statistically significant at the 95% level are shaded. (b) Same as (a) but for 200-hPa eddy streamfunction anomaly composite. Contour interval 1 × 10+6 m2 s−1. (c) Same as (a) but for 10–90-day filtered OLRA. (d) Same as (a) but for precipitation anomaly composite over the United States. Contour interval 0.3 mm day−1. Positive values are shaded.

Composites based on the box centered at 5°S, 125°E (Fig. 8) should be compared to composites keyed to California rainfall for category 2 dry events (Fig. 6). When the base point is shifted from 10°S to 5°N and from 110°E to 130°E, the pattern remains roughly the same and shows positive OLRA over California. Both OLRA composites (Figs. 6a and 8a) show an east–west three-cell pattern in the Tropics with enhanced convection in the western Pacific and positive anomalies in California. The positive OLRA (less rainfall) along thewest coast (Figs. 2a and 2d) is reproduced by the precipitation composite (Fig. 8d), which shows dryness in California, more rainfall in Washington, and less precipitation over the Southeast. The composite of OLRA in the intraseasonal band (Fig. 8c) shows the same east–west three-cell pattern in the Tropics and dryness in California, but the anomalies are weaker. The 200-hPa eddy streamfunction anomaly composite (Fig. 8b) corresponds well with the composite keyed to California precipitation (Fig. 6b). Both show a dipole straddling the equator and positive anomalies in Kamchatka and near California. In the Pacific North American sector, both patterns resemble the tropical Northern Hemisphere teleconnection pattern.

Fig. 8.

Same as Fig. 7 but for composites based on the OLRA averaged over a 10° × 10° box centered at 5°S, 125°E.

Fig. 8.

Same as Fig. 7 but for composites based on the OLRA averaged over a 10° × 10° box centered at 5°S, 125°E.

We conclude that the OLRA patterns associated with California precipitation can be reproduced in the composites keyed to convection in the Tropics. Large contributions to those anomalies are from fluctuations in the intraseasonal band. The tropical IO plays a role in modulating California rainfall for both wet and dry events with enhanced convection west of the dateline. Next, we present additional evidence to link the tropical IO to California precipitation by examining the evolution of category 2 wet and dry events.

2) Evolution of the category 2 wet events

At day −4, suppressed convection is located in the central Pacific and over the Indian Ocean and enhanced convection is located in the western Pacific centered at 130°E. The Pacific jet is located in the western Pacific, which is consistent with enhanced convection there (Fig. 9a). At day −2, the negative OLRA begin to shift eastward rapidly from 130° to 150°E (Figs. 9b–d). As enhanced convection propagates eastward into the central Pacific, the Pacific jet extends into the eastern Pacific and merges with the North American jet and rain starts to develop in California.

Fig. 9.

Low-passed OLRA (shaded) and 200-hPa zonal wind composite averaged over (a) day −4 to day −3, (b) day −2 to −1, (c) day 0 to 1, and (d) day 2 to 3 for wet category 2 events. Unit for OLRA: W m−2. Contour interval for 200-hPa zonal wind 30 m s−1.

Fig. 9.

Low-passed OLRA (shaded) and 200-hPa zonal wind composite averaged over (a) day −4 to day −3, (b) day −2 to −1, (c) day 0 to 1, and (d) day 2 to 3 for wet category 2 events. Unit for OLRA: W m−2. Contour interval for 200-hPa zonal wind 30 m s−1.

The evolution of wet events can be illustrated by the time longitude section of OLRA, 200-hPa streamfunction, and 200-hPa wind anomalies (Fig. 10). Enhanced tropical convection is located in the western Pacific before day −4 and the Pacific jet is located also in the western Pacific as indicated by the positive 200-hPa wind anomalies (Fig. 10c). The rapid eastward propagation of convection starts at day −3. The negative OLRA reach 150°E at day 0 and at the same time, a wave train extending from the area of enhanced convection to the west coast of the United States (Fig. 10d) starts to form. A trough establishes itself near California at day −2 and reaches a maximum at day 5. After onset, westerlies extend to the west coast and reach a maximum at day 7 while precipitation in California persists. After onset, most contributions to negative OLRA centered at 150°E are from fluctuations in the intraseasonal band (Fig. 10b).

Fig. 10.

Time and longitude cross section from day −10 to day 10 for the wet category 2 composite. (a) OLRA averaged from 5°S to 10°N. Contour interval 3 W m−2. Zero contours are omitted. Negative values are shaded. (b) Same as (a) but for 10–90-day filtered OLRA. (c) Same as (a) but for 200-hPa zonal wind anomalies averaged from 30° to 35°N. Contour interval 3 m s−1. Positive values are shaded. (d) Same as (a) but for 200-hPa eddy streamfunction anomalies averaged from 35° to 45°N. Contour interval 2 × 10+6 m2 s−1. Contours −1 and 1 × 10+6 m2 s−1 are added.

Fig. 10.

Time and longitude cross section from day −10 to day 10 for the wet category 2 composite. (a) OLRA averaged from 5°S to 10°N. Contour interval 3 W m−2. Zero contours are omitted. Negative values are shaded. (b) Same as (a) but for 10–90-day filtered OLRA. (c) Same as (a) but for 200-hPa zonal wind anomalies averaged from 30° to 35°N. Contour interval 3 m s−1. Positive values are shaded. (d) Same as (a) but for 200-hPa eddy streamfunction anomalies averaged from 35° to 45°N. Contour interval 2 × 10+6 m2 s−1. Contours −1 and 1 × 10+6 m2 s−1 are added.

3) Evolution of category 2 dry events

These events are likely to occur when tropical convection is suppressed in the central Pacific. The convection pattern with positive OLRA in the central Pacific and negative OLRA over the Indian Ocean is established about 10 days before onset (Fig. 11a). The subtropical Pacific jet extends to 130°W near the west coast and the North American jet is located at its normal position. When negative OLRA propagate eastward from the Indian Ocean to the western Pacific, positive OLRA also slowly shift eastward. Suppressed convection in the central Pacific weakens the local Hadley circulation. The weakening of upper-level convergence in the subtropics creates a favorable condition for cloud bands to form in the North Pacific as indicated by negative OLRA in the subtropics centered at 25°N. At day −5, the Pacific jet starts to shift westward. The storm tracks follow the jet and move westward and that leaves California dry. After onset, the North America jet extends further into the Atlantic.

Fig. 11.

Low-passed OLRA (shaded) and 200-hPa zonal wind composite at (a) day −10, (b) day −5, (c) day 0, and (d) day 5 for dry category 2 events. Unit for OLRA: W m−2. Contour interval for 200-hPa zonal wind: 30 m s−1.

Fig. 11.

Low-passed OLRA (shaded) and 200-hPa zonal wind composite at (a) day −10, (b) day −5, (c) day 0, and (d) day 5 for dry category 2 events. Unit for OLRA: W m−2. Contour interval for 200-hPa zonal wind: 30 m s−1.

The evolution of dry events can be illustrated by the time longitude section of OLRA, 200-hPa eddy streamfunction, 200-hPa wind anomalies (Fig. 12). Enhanced tropical convection (negative OLRA) propagates eastward from the Indian Ocean to the western Pacific (Fig. 12a) while suppressed OLRA are located in the central Pacific east of the dateline. Large contributions are from fluctuations from the 10–90-day band (Fig. 12b). The negative zonal wind anomalies east of the dateline signal the retraction of the subtropical jet (Fig. 12c). The negative OLRA reach 100°–125°E at day −2 and at the same time, a wave train extending from the area of enhanced convection to the west coast of the United States (Fig. 12d) starts to form. The ridge establishes itself near California at day −2 and reaches a maximum at day 2. After onset, the Pacific jet remains near the dateline and drought in California persists.

Fig. 12.

Same as Fig. 10 but for dry category 2 composites.

Fig. 12.

Same as Fig. 10 but for dry category 2 composites.

7. Conclusions

The NCEP–NCAR reanalysis together with OLRA and gridded daily precipitation over North America was used to examine tropical influences on California precipitation. The precipitation anomaly pattern over the Pacific near the west coast shows a north–south three-cell pattern. Heavy precipitation in California is accompanied by dryness over Washington, British Columbia, and the southeast coast of Alaska and suppressed precipitation in the eastern Pacific. The seesaw between rainfall in California and the Pacific Northwest is related to the transport of moisture. During California wet events, the southern branch of moisture transport strengthens and brings moisture to California. During dry events, the enhanced northern branch of moisture flux transport brings more moisture to the Pacific Northwest and leaves California dry.

The seesaw between California rainfall and rainfallanomalies in the eastern tropical Pacific just north of the equator has dynamic support. The enhanced precipitation in the eastern Pacific just north of the ITCZ generates strong rising motion. The associated strong sinking motion over California and suppressed low-level moisture convergence favor dry conditions in California. Conversely, suppressed rainfall in the eastern Pacific is associated with wet events.

As suggested by Schonher and Nicholson (1989), Cayman and Redmond (1994), and many others, we found that dry events are favored during cold ENSO events, but not all warm ENSO events produce heavy rain in California. Local influences are as important as the remote forcing in determining California precipitation. California has above-normal rainfall when the precipitation in the subtropical eastern Pacific is suppressed and convection is enhanced in the central Pacific.

In addition to ENSO, precipitation in California is also modulated by the tropical IO. Enhanced convection in the central Pacific centered near 150°E is related to wet events and enhanced convection in the western Pacific near 120°E is linked to dry events.

Acknowledgments

This investigation was partially supported by Interagency Agreement S-41367-F under the authority of NASA/GSFC, and by the NOAA Office of Global Programs under the GEWEX Continental Scale International Project.

REFERENCES

REFERENCES
Blackmon, M. L., 1976: A climatological spectral study of the 500 mb height of the Northern Hemisphere. J. Atmos. Sci., 33, 1607–1623.
Cayan, D. R., and D. H. Peterson, 1989: The influence of North Pacific atmospheric circulation on streamflow in the west. PACLIM. AGU Monogr., No. 55, Amer. Geophys. Union, 375–397.
——, and R. H. Webb, 1992: El Niño/Southern Oscillation and streamflow in the western United States 29–68. Historical and Paleoclimatic Aspects of the Southern Oscillation H. F. Diaz and V. Markgraf, Eds., Cambridge University Press, 412 pp.
——, and K. T. Redmond, 1994: ENSO influences on atmospheric circulation and precipitation in the western United States. Proc. 10th Annual Pacific Climate (PACLIM) Workshop, Pacific Grove, CA, California Department of Water Resources, 5–26.
Dole, R. M., and N. D. Gordon, 1983: Persistent anomalies of the extratropical Northern Hemisphere wintertime circulation: Geographical distribution and regional persistent characteristics. Mon. Wea. Rev., 111, 1567–1587.
Fu, C., H. F. Diaz, and J. O. Fletcher, 1986: Characteristics of the response of SST in the central Pacific associated with warm episodes of the Southern Oscillation. Mon. Wea. Rev., 114, 1716–1738.
Higgins, R. W., J. E. Janowiak, and Y. P. Yao, 1996a: A gridded hourly precipitation data base for the United States (1963–1993). NCEP/Climate Prediction Center Atlas No. 1, NCEP/NWS/NOAA, 47 pp. [Available from NCEP/NWS/NOAA, Washington, DC 20233.]
.
——, K. C. Mo, and S. D. Schubert, 1996b: The moisture budget of the central United States in spring as evaluated in the NCEP/NCAR and the NASA/DAO reanalyses. Mon. Wea. Rev., 124, 939–963.
Janowiak, J. E., 1992: Tropical rainfall: A comparison of satellite derived rainfall estimates with model precipitation forecasts, climatologies and observations. Mon. Wea. Rev., 120, 448–461.
Kahya, E., and J. A. Dracup, 1994: The influence of type 1 El Niño and La Niña events on streamflows in the Pacific Southwest of the United States. J. Climate, 7, 965–976.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Meteor. Soc., 77, 437–471.
Lanzante, J. R., and R. P. Harnack, 1982: Specification of United States summer season precipitation. Mon. Wea. Rev., 110, 1843–1850.
Lau, K. M., and P. H. Chan, 1985: Aspects of the 40–50 day oscillation during the northern winter as inferred from outgoing long wave radiation. Mon. Wea. Rev., 113, 1889–1909.
Liebmann, B., and C. A. Smith, 1996: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Amer. Meteor. Soc., 77, 1275–1277.
Livezey, R. E., and K. C. Mo, 1987: Tropical–extratropical teleconnections during the Northern Hemisphere winter. Part II: Relationships between monthly mean Northern Hemisphere circulation patterns and proxies for tropical convection. Mon. Wea. Rev., 115, 3115–3123.
Mo, K. C., and R. W. Higgins, 1996: Large-scale atmospheric moisture transport as evaluated in the NCEP/NCAR and the NASA/DAO reanalyses. J. Climate, 9, 1531–1545.
——, J. N. Paegle, and R. W. Higgins, 1997: Atmospheric processesassociated with summer floods and droughts in the central United States. J. Climate, 10, 3028–3046.
Paegle, J. N., and K. C. Mo, 1997: Alternating wet and dry conditions over South America during summer. Mon. Wea. Rev, 125, 279–291.
Papoulis, A., 1973: Minimum bias windows for high resolution spectral estimates. IEEE. Trans. Info. Theory, 19, 9–12
.
Richman, M. B., and P. J. Lamb, 1985: Climate pattern analysis of three and seven day summer rainfall in the central United States:Some meteorological considerations and a regionalization. J. Climate Appl. Meteor., 24, 1325–1343.
Ropelewski, C. F., and M. S. Halpert, 1986: North American precipitation and temperature patterns associated with the El Niño/Southern Oscillation (ENSO). Mon. Wea. Rev., 114, 2352–2362.
Schonher, T., and S. E. Nicholson, 1989: The relationship between California rainfall and the ENSO events. J. Climate, 2, 1258–1269.
Trenberth, K. E., and C. J. Guillemot, 1995: Evaluation of the atmospheric moisture budget as seen from analyses. J. Climate, 8, 2255–2272.
Wallace, J. M., and D. S. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon. Wea. Rev., 109, 784–812.
Xie, P., and P. A. Arkin, 1996: Analyses of global monthly precipitation using gauge, observations, satellite estimates and numerical model predictions. J. Climate, 9, 840–858.

Footnotes

Corresponding author address: Kingtse C. Mo, Climate Prediction Center, NCEP/NWS/NOAA W/NP52, 4700 Silver Hill Rd., Stop 9910, Washington, DC 20233-9910.