Abstract

The relative contributions to mean global atmospheric moisture transport by both the time-mean circulation and by synoptic and low-frequency (periods greater than 10 days) anomalies are evaluated from the vertically integrated atmospheric moisture budget based on 40 yr of “chi corrected” NCEP–NCAR reanalysis data. In the extratropics, while the time-mean circulation primarily moves moisture zonally within ocean basins, low-frequency and synoptic anomalies drive much of the mean moisture transport both from ocean to land and toward the poles. In particular, during the cool-season low-frequency variability is the largest contributor to mean moisture transport into southwestern North America, Europe, and Australia. While some low-frequency transport originates in low latitudes, much is of extratropical origin due to large-scale atmospheric anomalies that extract moisture from the northeast Pacific and Atlantic Oceans. Low-frequency variability is also integral to the Arctic (latitudes > 70°N) mean moisture budget, especially during summer, when it drives mean poleward transport from relatively wet high-latitude continental regions. Synoptic variability drives about half of the mean poleward moisture transport in the midlatitudes of both hemispheres, consistent with simple “lateral mixing” arguments. Extratropical atmospheric transport is also particularly focused within “atmospheric rivers” (ARs), relatively narrow poleward-moving moisture plumes associated with frontal dynamics. AR moisture transport, defined by compositing fluxes over those locations and times where column-integrated water vapor and poleward low-level wind anomalies are both positive, represents most of the total extratropical meridional moisture transport. These results suggest that understanding potential anthropogenic changes in the earth ’s hydrological cycle may require understanding corresponding changes in atmospheric variability, especially on low-frequency time scales.

1. Introduction

Perhaps the most striking feature in a satellite loop of the earth is that the atmosphere transports water across great distances. While this transport appears to occur more frequently in certain regions, it is not steady; rather, it is characterized by numerous transient features of many scales. Thus, to understand atmospheric moisture transport, including the dual role it plays in the global energy cycle and as the source of water over the continents, there is a need to understand how atmospheric variability on different time scales acts to transport moisture and, in turn, is affected by it (e.g., Schneider et al. 2010; Trenberth 2011). This is true not only for variations in moisture transport, including extreme precipitation events, but also for the impact of atmospheric variability on the climatological mean moisture transport that is the subject of this paper.

It is well known that transient eddies are a critical part of poleward mean moisture transport in the extratropics (e.g., Peixoto and Oort 1992). More recently, it has been suggested that virtually all extratropical moisture transport is focused within long, relatively narrow bands sometimes called “atmospheric rivers” (ARs; Zhu and Newell 1998, hereafter ZN; Ralph et al. 2004; Neiman et al. 2008b, hereafter N08). ZN suggested that moisture transport is predominantly confined to these ARs, so that at any given time and at any given latitude about 90% of the meridional moisture transport occurs within only 10% of the zonal band. ARs are particularly striking in column-integrated water vapor (IWV), such as is measured by the Special Sensor Microwave Imager (SSM/I), and at times they extend from deep in the tropics to midlatitudes (Ralph et al. 2011). The extent to which such ARs represent the transport of moisture from the tropics to the extratropics has been a matter of some debate (e.g., Bao et al. 2006; Knippertz and Wernli 2010), although recent research aircraft observations have confirmed that transport from the tropics can occur (Ralph et al. 2011). Studies have shown that the IWV bands generally are regions of strong surface convergence, and that their leading edges typically correspond to the strong moist low-level jet sometime called the “moist conveyor belt” associated with fronts (Bao et al. 2006; Knippertz and Wernli 2010). Additionally, observational case studies (Ralph et al. 2004, 2005, 2011; Neiman et al. 2008a), composites of many aircraft-observed events (Ralph et al. 2005), and statistical comparison of 8 yr of reanalysis against SSM/I observations (Ralph et al. 2006; N08) show that the IWV bands can correspond to regions of pronounced moisture flux, that is, atmospheric rivers.

Two complementary approaches have been used to investigate moisture transport. Given the episodic nature of the IWV bands, it seems natural to use a Lagrangian framework (e.g., Stohl and James 2005; Bao et al. 2006; Eckhardt et al. 2004; Dirmeyer and Brubaker 2007; Knippertz and Wernli 2010; Gimeno et al. 2010; Drumond et al. 2011) and follow the trajectories of individual moist air masses, either forward from many different starting locations to determine where the moisture ultimately goes or backward starting from specified locations and/or precipitation events to find relevant sources. For climate studies the analysis can be computationally expensive, and the trajectory model can be sensitive to errors in the input atmospheric fields as well as errors in the parameterizations and represented dynamics. Also, water vapor is not entirely a passive tracer, so the types of trajectories that can be considered are either limited, especially in their duration, or some assumption must be made to keep track of water phase changes.

For our purposes, it is more straightforward to separate the effects of different time scales of variability in a Eulerian rather than Lagrangian analysis. Eulerian analyses are well suited for determining the relative magnitudes of different processes. Past Eulerian analyses have generally used global fields generated by state-of-the-art four-dimensional data assimilation systems to evaluate the moisture budget, defined as

 
formula

where is IWV (also sometimes called “precipitable water”), E is evaporation from the surface, P is precipitation, and Q is the vertically integrated moisture flux (e.g., Peixoto and Oort 1992). We could determine Q as a residual from (1) using independent estimates of precipitation and evaporation, including those output by the GCMs used to produce the “first guess” fields for data assimilation. However, the atmospheric moisture budget determined from these fields does not balance observed streamflow runoff even on longer time scales (e.g., Betts et al. 1999; Lenters et al. 2000; Roads and Betts 2000). Alternatively, Q can be computed using the analyzed wind and humidity fields. This approach has been extensively applied, initially using operational analyses (Trenberth 1991; Roads et al. 1994; Trenberth and Guillemot 1995; Wang and Paegle 1996) but more recently using reanalyses (Higgins et al. 1996; Mo and Higgins 1996; Higgins et al. 1997; Gutowski et al. 1997; Min and Schubert 1997; Trenberth and Guillemot 1995, 1998; Betts et al. 1999; Roads et al. 2002; Mo et al. 2005; Schneider et al. 2006; Trenberth et al. 2007, 2011; Pauluis et al. 2011; Shaw and Pauluis 2012; and many others). Globally, and particularly in the extratropics, there is much better agreement among different reanalyses for PE computed as a residual from (1) than for PE computed from the reanalysis estimates of P and E (e.g., Trenberth et al. 2011). This gives us some confidence in estimates of moisture transport and its divergence as generated from reanalysis humidity and wind fields, and this is the approach we take in this paper.

A few past studies have divided atmospheric moisture transport into contributions from the transient, zonal mean, and stationary wave portions of the circulation (e.g., Peixoto and Oort 1992; Shaw and Pauluis 2012). However, most recent studies of variability on “low frequency” (LF) time scales (e.g., intraseasonal to interannual) typically define anomalies as departures from the time-averaged atmospheric state, since storm-track and climate dynamics in the troposphere are both strongly influenced by zonal and meridional asymmetries of the basic state (e.g., Blackmon et al. 1977; Simmons et al. 1983; Borges and Sardeshmukh 1995; Whitaker and Sardeshmukh 1998; Winkler et al. 2001; Chang et al. 2002; and many others). In this study, we investigate the separate contributions of synoptic and low-frequency anomalies, defined as time-varying departures from the seasonally varying basic state and split into high-frequency (periods < 10 days) and low-frequency (periods > 10 days) components, by determining their relative importance in the seasonally varying climatological mean moisture budget determined from (1). This approach is laid out in section 2 along with a description of the 40-yr-long dataset. Results are in section 3, where we find that despite the dominance of moisture transport by the mean circulation over the oceans, synoptic and LF time scales play critical roles in both meridional and ocean-to-land time-mean moisture transport. The contribution of atmospheric rivers to moisture transport is assessed in section 4, and a closer focus on extratropical LF moisture transport is in section 5. Concluding remarks are in section 6.

2. Data and analysis

In sigma coordinates, the vertically integrated moisture flux in (1) is , where q is specific humidity, ps is surface pressure, v is the horizontal wind vector, and brackets indicate the vertical integral in sigma coordinates, from σ = 1 to σ = 0 (e.g., Trenberth and Guillemot 1995). The vertical integral is most accurately done on the original data levels (Trenberth 1991), which for the dataset used in this paper are sigma coordinates.

We define and similarly for all other variables, where overbars indicate the seasonally varying climatological mean and superscripts indicate low-frequency (LF) and synoptic (s) time-scale anomalies. Anomalies were defined by removing each variable ’s annual cycle, determined from the first three harmonics of the 4-times-daily dataset, at each grid point. LF anomalies are determined from a 121-point Lanczos filter that passes periods greater than 10 days, and synoptic anomalies are the residuals representing periods less than 10 days; this frequency cutoff was chosen since it is common to a great many studies of LF variability dating back to Blackmon et al. (1977). Applying these definitions to (1), we obtain the following seasonally varying mean moisture budget:

 
formula

where the mean moisture transport is

 
formula
 
formula

These terms were computed for each month of the year. The latter three terms within each curly bracket in (3) that depend on ps anomalies are included in our computations of and , but they are generally negligible. Thus, for the remainder of this paper, we label the terms on the right-hand side of (4) as

 
formula

We stress that these are shorthand expressions for mean moisture transports by the time-mean flow, by LF anomalies, and by synoptic anomalies.

The moisture budget was computed for 40 yr of National Centers for Environmental Prediction (NCEP)–National Center for Atmospheric Research (NCAR) reanalysis data covering the period 1968–2007, with wind fields adjusted toward momentum and mass balance using the improved iterative solution of the “χ problem” (Sardeshmukh 1993) discussed in the  appendix. All moisture budget calculations were repeated using NCEP–Department of Energy (DOE) Global Reanalysis 2 (NCEP-2) data for 1979–2005; for the common period, the results have only minor quantitative differences, so for brevity they are not displayed here. After their calculation, all moisture flux divergence fields were smoothed for display purposes using the Sardeshmukh and Hoskins (1984, hereafter SH) spatial filter with n = 42 and r = 2.

3. Seasonal variation of the atmospheric moisture budget

a. Winter and summer global moisture transport

Figures 1 and 2 show the results of (4) for December–February (DJF) and June–August (JJA) 1968–2007. Each figure shows the moisture transport terms , , and (vectors) and corresponding moisture flux divergences (shading). The top-left panel shows and , the mean atmospheric moisture transport and moisture flux divergence; that is, the sum of the three remaining panels. In these two figures, is nearly equal to , since the tendency of the seasonally varying climatology is quite small near the solstices.

Fig. 1.

Terms in the mean atmospheric water budget (2) for DJF of 1968–2007. Vertically integrated moisture flux divergence (shading) and vertically integrated moisture flux (vectors) for (a) total, (b) mean, (c) LF, and (d) synoptic terms from (4). Moisture flux vectors are scaled by (top) 300 and (bottom) 30 kg m−1 s−1; the moisture flux divergence contour level, however, is the same in all four panels. For display purposes, moisture flux divergences are smoothed with the SH filter (see text).

Fig. 1.

Terms in the mean atmospheric water budget (2) for DJF of 1968–2007. Vertically integrated moisture flux divergence (shading) and vertically integrated moisture flux (vectors) for (a) total, (b) mean, (c) LF, and (d) synoptic terms from (4). Moisture flux vectors are scaled by (top) 300 and (bottom) 30 kg m−1 s−1; the moisture flux divergence contour level, however, is the same in all four panels. For display purposes, moisture flux divergences are smoothed with the SH filter (see text).

Fig. 2.

As in Fig. 1, but for JJA of 1968–2007.

Fig. 2.

As in Fig. 1, but for JJA of 1968–2007.

The transport by the mean dominates the mean moisture transport (note that the vectors representing both and are scaled by a factor of 10 larger than for the other two terms). This is also true for the transition seasons (not shown). The gross features of the mean transport are familiar from other studies, showing broad areas of moisture sources (positive ) within the subtropics and sinks over the continents, the midlatitude storm tracks, and the warm pool and monsoon regions of the tropics. The transport by the mean during DJF in the Northern Hemisphere extratropics is mostly zonal and only somewhat poleward from the western to the eastern margins of the North Pacific and North Atlantic Ocean basins. There is also strong zonal transport by the mean in the Southern Hemisphere extratropics, but it is sufficiently uniform that its associated moisture divergence is relatively small. Extratropical zonal transport is stronger in both hemispheres during DJF than during JJA. The net transport by winds swirling around the mean subtropical highs in the summer hemispheres is primarily westward, but there is also some higher-latitude poleward transport.

Figures 1 and 2 might give a somewhat misleading picture of the water cycle because a large fraction of the atmospheric moisture transport essentially moves water zonally from one part of the ocean to another (which does have important implications for ocean dynamics by changing surface salinity; see, e.g., Huang 1993; Delcroix et al. 1996). While this “ocean to ocean” moisture transport is dominated by , both ocean-to-land transport and meridional transport are not, which can initially be seen by close inspection of Figs. 1 and 2. Also, even though and are generally much weaker than , their corresponding flux divergences are comparable in many regions (note that in Figs. 1 and 2, the contour interval of flux divergence is the same in all panels), so that anomaly moisture transport is important to the overall moisture budget.

b. Ocean-to-land moisture transport

The importance of and for ocean-to-land moisture transport is shown in Figs. 3a–e, which displays the inflow of moisture into several different land regions (determined from the areal average of each corresponding moisture convergence term within each region) as a function of the seasonal cycle (these computations have also been made for the NCEP-2 dataset with only minor differences for the common period; not shown). In the global average (not shown), , , and drive about 77%, 12%, and 11%, respectively, of the annual mean ocean-to-land transport. This result, however, is dominated by the tropics (Fig. 3a), where the mean trades transport moisture from the ocean into the tropical convergence zones, including those over land, with LF anomalies (and, to a lesser extent, synoptic anomalies) acting to transport some moisture out of the convergence zones back to the ocean. In contrast, in the extratropics (Figs. 3b and 3c) the three terms have roughly equal importance for annual mean ocean-to-land transport, since during the cool season moisture is mainly transported by both and , more than offsetting the usually dominant role of during summer (e.g., JJA in the Northern Hemisphere and DJF in the Southern Hemisphere).

Fig. 3.

(left) Seasonal cycle of moisture transport from ocean to land in the (a) tropics, (b) Northern Hemisphere extratropics, and (c) Southern Hemisphere extratropics, where the tropical–extratropical boundaries are set as 20°N and 20°S. (right) Seasonal cycle of moisture transport into the specified regions of (d) North America, (e) Europe (where the eastern boundary is set as 60°E), and (f) the Arctic (defined as the region north of 70°N). The last pair of bars in each panel shows the annual mean terms. Mean moisture flux (transport) driven by the mean, synoptic, and LF anomalies into each region is determined by the areal average of the vertically integrated moisture flux convergence over each region. Mean precipitable water () tendency is also averaged in the region and multiplied by −1. Thus, the sign of all terms is chosen so that their sum (determined from the stacked bars on the left of each pair) is equal to PE also averaged in each region (blue bar on the right of each pair).

Fig. 3.

(left) Seasonal cycle of moisture transport from ocean to land in the (a) tropics, (b) Northern Hemisphere extratropics, and (c) Southern Hemisphere extratropics, where the tropical–extratropical boundaries are set as 20°N and 20°S. (right) Seasonal cycle of moisture transport into the specified regions of (d) North America, (e) Europe (where the eastern boundary is set as 60°E), and (f) the Arctic (defined as the region north of 70°N). The last pair of bars in each panel shows the annual mean terms. Mean moisture flux (transport) driven by the mean, synoptic, and LF anomalies into each region is determined by the areal average of the vertically integrated moisture flux convergence over each region. Mean precipitable water () tendency is also averaged in the region and multiplied by −1. Thus, the sign of all terms is chosen so that their sum (determined from the stacked bars on the left of each pair) is equal to PE also averaged in each region (blue bar on the right of each pair).

During DJF, the pattern of is very different from that of , and the corresponding “source” regions are also different (cf. Figs. 1c and 1d). Notably, in both the eastern extratropical Atlantic and Pacific basins there are LF source regions that export moisture, both eastward and westward as well as poleward. Although it might seem small in Figs. 1 and 2, LF transport is the difference between a moisture deficit and surplus in key areas such as the North American southwest and Europe (Fig. 3e) in DJF, and southern Australia and New Zealand in JJA, especially since in all these areas the mean flow acts to remove moisture. Averaged over the entire North American continent, the synoptic transport is slightly larger than LF transport (Fig. 3d); however, the relative importance of these two terms varies between the western half of North America, where LF transport from sources in both the extratropical and subtropical east Pacific is greater (more so for the northwest and southwest portions of the continent, respectively), and the eastern half of North America, where the synoptic transport is greater. For Europe, both the Atlantic and Mediterranean act as LF sources (see also Gimeno et al. 2010). Note also that both the LF and synoptic transports substantially broaden the moisture sinks on the poleward sides of the subtropical highs during JJA.

c. Meridional moisture transport

The well-known importance of transients in driving mean meridional moisture transport (e.g., Peixoto and Oort 1992) is evident in Fig. 4, which shows the zonal average of each term contributing to , the meridional component of , as a function of the seasonal cycle (Figs. 4a–d) and for the annual mean (Fig. 4e). In the tropics and subtropics, transport by the mean circulation dominates the meridional transport, primarily equatorward but with a cross-equatorial component toward the summer hemisphere near the solstices (see also Schneider et al. 2006). In the extratropics the synoptic transport is the largest term, especially in the Southern Hemisphere. As we might expect, the synoptic transport is maximized within the storm tracks over the western part of both Northern Hemisphere oceans (Fig. 1d) where synoptic variability is strongest (e.g., Chang et al. 2002; see also Fig. 10b), with the maxima moving poleward with the storm tracks in summer (Fig. 2d). Note that in the extratropics, the moisture source for synoptic transport lies poleward of the source (cf. Figs. 1b and 1d). Also, synoptic transport is almost entirely meridional, with its small zonal component making little contribution to flux divergence. Meridional LF transport is only about half of the synoptic transport but acts over a broader latitude range, with a somewhat greater effect in the Pacific sector (not shown, but see Figs. 1c and 2c) than for the zonal mean.

Fig. 4.

Seasonal cycle of terms contributing to vertically integrated, zonally averaged mean meridional moisture flux : (a) mean transport, (b) transport by the mean, (c) LF transport, and (d) synoptic transport. Contour interval is 2.5 kg m−1 s−1; positive is northward and negative is southward. (e) Annual mean vertically integrated, zonally averaged meridional moisture flux .

Fig. 4.

Seasonal cycle of terms contributing to vertically integrated, zonally averaged mean meridional moisture flux : (a) mean transport, (b) transport by the mean, (c) LF transport, and (d) synoptic transport. Contour interval is 2.5 kg m−1 s−1; positive is northward and negative is southward. (e) Annual mean vertically integrated, zonally averaged meridional moisture flux .

Moisture transport into the polar regions can be determined by the value of the zonally averaged meridional transport at 70°N and 70°S in Fig. 4. Mean transport into the Arctic (Fig. 3f) peaks during summer, in agreement with earlier studies (e.g., Serreze et al. 2006, 2007). LF anomalies drive close to two-thirds of this transport every month of the year. In the winter the LF transport occurs primarily over the Atlantic, but in the summer it is dominated by poleward transport from the large landmasses toward the Arctic Ocean (cf. Figs. 1c and 2c). Much of this transport occurs from Eurasia, consistent with an increased frequency of blocking there during summer (e.g., Tyrlis and Hoskins 2008; Dole et al. 2011) as well as a pronounced summertime maximum over northern Eurasia in variance (not shown). We find that LF moisture anomalies are in phase with LF meridional wind anomalies related to blocks (not shown), consistent with the poleward moisture transport occurring as anomalous southerlies (northerlies) to the west (east) of the block transport anomalously moist (dry) air from the continent (polar region). In contrast, synoptic and LF transports into the Antarctic are about equally important (not shown), and these have spring and fall maxima, consistent with the Southern Hemisphere semiannual oscillation (van Loon 1967).

d. Analysis of transport terms

It is instructive to separate moisture flux divergence into the moisture advection and divergence terms, and , respectively, computed separately from the mean and from the LF and synoptic anomalies, shown in Fig. 5 for DJF. In the extratropics, strong mean zonal flow advects continental dry air over the warm western boundary currents, which act as moisture sources for the atmosphere. In the central and northeastern extratropical ocean basins, moisture sinks are balanced by both the northeastward advection of moist air over increasingly cooler oceans and particularly up over much drier continents, and by strong moisture convergence due to the deceleration of the winds in the jet exit region. In the storm tracks, where rising (subsiding) motion generally occurs in anomalously wet (dry) regions, the convergence term acts as a synoptic moisture sink, while the advection term serves primarily as the synoptic moisture source. In contrast, LF moisture sources and sinks are driven almost entirely by advection alone since the convergence term is negligible, consistent with the more barotropic nature of large-scale LF anomalies.

Fig. 5.

Contributors to moisture flux divergence for DJF of 1968–2007: (left) divergence terms , , and ; and (right) advection terms , , and . For display purposes, all terms are smoothed with the SH filter (see text).

Fig. 5.

Contributors to moisture flux divergence for DJF of 1968–2007: (left) divergence terms , , and ; and (right) advection terms , , and . For display purposes, all terms are smoothed with the SH filter (see text).

The time-scale dependence of LF transport is examined by applying additional time filters. As in section 2, we define , and similarly for all other variables, where the superscripts represent the bandpass interval (in days) determined from additional applications of the Lanczos filter, with 90+ representing a filter that passes all periods longer than 90 days. The relative importance of different bands during DJF varies by location, as shown in Fig. 6. Transport by the 10–30-day band represents more than half the LF transport into the eastern half of North America but under one-third over the North Atlantic and Europe, where longer time-scale (>90 days) anomalies drive almost half. The strongest moisture convergence occurs in the 10–30-day band along the eastern, southern, and western boundaries of the continental United States, associated with strong northward or northeastward transports. Most of the LF transport in the Southern Hemisphere is in the 10–30-day band, with the notable exception of southern Australia. However, the transport out of the South Pacific convergence zone (SPCZ) gains an additional eastward component with increasing time scale, so that its moisture flux divergence is about the same for each band (this occurs year-round; not shown). Clearly much of the result in Fig. 6 is likely not indicative of changing moisture transport dynamics with time scale but instead is consistent with time filtering of a red-noise process. This is in contrast to the differences between the LF and synoptic panels in Fig. 1, where the moisture sources and sinks for the LF and synoptic transports are so different that the LF transport seems unlikely to be merely a low-frequency residual of the white noise of individual extratropical storms.

Fig. 6.

Moisture transport by bandpass anomalies for DJF of 1968–2007: vertically integrated moisture flux divergence (shading) and corresponding moisture fluxes (vectors). Note that the moisture flux divergence contour level is reduced (relative to Fig. 1) to 0.125 mm day−1 and the moisture flux vectors are scaled by 10 kg m−1 s−1. For display purposes, moisture flux divergences are smoothed with the SH filter (see text).

Fig. 6.

Moisture transport by bandpass anomalies for DJF of 1968–2007: vertically integrated moisture flux divergence (shading) and corresponding moisture fluxes (vectors). Note that the moisture flux divergence contour level is reduced (relative to Fig. 1) to 0.125 mm day−1 and the moisture flux vectors are scaled by 10 kg m−1 s−1. For display purposes, moisture flux divergences are smoothed with the SH filter (see text).

4. Impact of “atmospheric rivers” on the atmospheric moisture transport

The statistical analysis described in the preceding section, of course, yields a picture that is representative of the net effect of all individual events. In this section we examine this more closely, also considering the impact of atmospheric rivers as described, for example, by ZN and N08. As an example, Fig. 7 shows a time–longitude diagram of the 4-times-daily Q and at 35° (left) and 45°N (right) for DJF 2001/02 and 2002/03. For comparison, Fig. 8 repeats the left panels of Fig. 7 but uses data in which either the synoptic components or the LF components are removed before Q and are computed. Even over these relatively short periods, we can see elements of the different time scales for transport shown in Fig. 1. In the central Pacific, moisture transport is clearly dominated by synoptic events, with pronounced poleward transport occurring in fairly narrow bands that occur with some regularity and that propagate rapidly eastward. There is almost no corresponding equatorward transport, and in particular there appears to be almost no poleward moisture transport that results from dry air advected equatorward. In contrast, near the midlatitude coasts (around 120° and 5°W), the LF transport component is more evident with much longer periods of strong and weak transport, especially inland, and even some instances of equatorward transport (see also Fig. 8). At higher latitudes in both ocean basins, the transport almost always occurs within a fairly narrow longitude range and is much more persistent overall.

Fig. 7.

Hovmöller of 4-times-daily total vertically integrated moisture flux Q (vectors, scaled by 600 kg m−1 s−1) and vertically IWV (shading, contour interval 5 mm) at (a) 35° and (b) 45°N for the winters (DJF) of 2001/02 and 2002/03. Note that vectors show the local direction of the flux.

Fig. 7.

Hovmöller of 4-times-daily total vertically integrated moisture flux Q (vectors, scaled by 600 kg m−1 s−1) and vertically IWV (shading, contour interval 5 mm) at (a) 35° and (b) 45°N for the winters (DJF) of 2001/02 and 2002/03. Note that vectors show the local direction of the flux.

Fig. 8.

As in Fig. 7a, but for Q and determined from (left) data with synoptic components removed or (right) LF components removed.

Fig. 8.

As in Fig. 7a, but for Q and determined from (left) data with synoptic components removed or (right) LF components removed.

Because at any given time much of the transport is located within fairly narrow spatial bands, ZN suggested that the bands represent atmospheric rivers, which they defined as filament-like structures of moisture flux representing most of the global total moisture transport. They categorized these regions by finding all locations where the magnitude of Q, |Q|, was relatively higher than its zonal mean value. Specifically, their algorithm determined that a river existed wherever and whenever |Q| ≥ |Qmean| + 0.3(|Qmax| − |Qmean|), where Qmean is the zonal mean Q and Qmax is the longitudinal maximum, both of which are functions of latitude and time.

To gain a comprehensive picture of the effect of ARs on the atmospheric moisture transport, we composited Q in those regions and times in our 40-yr dataset where AR conditions occur, using the ZN definition, and also determined the frequency of AR condition occurrence worldwide. The results (Fig. 9a) confirm that, as ZN suggested from much more limited data, the flux associated with atmospheric rivers defined in this way represents a large portion of the total moisture flux field, and virtually all of the extratropical meridional transport (not shown). However, a comparison of Fig. 9d to Fig. 1a shows that the AR composite takes into account neither transport by the mean subtropical highs in the Southern Hemisphere (and similarly for the Northern Hemisphere during summer; not shown) nor the substantial zonal transport that remains in the extratropical jets of both hemispheres.

Fig. 9.

DJF 4-times-daily moisture flux composited by different AR criteria: (a) ZN criterion, (b) N08 criterion, and (c) “positive anomalous plus positive anomalous ” criterion. Corresponding difference between (d)–(f) total moisture flux (Fig. 1a) and (a)–(c) composite flux. Flux vectors are scaled by 300 kg m−1 s−1. Shading in panels indicates the frequency of occurrence of (a)–(c) AR and (d)–(f) non-AR conditions.

Fig. 9.

DJF 4-times-daily moisture flux composited by different AR criteria: (a) ZN criterion, (b) N08 criterion, and (c) “positive anomalous plus positive anomalous ” criterion. Corresponding difference between (d)–(f) total moisture flux (Fig. 1a) and (a)–(c) composite flux. Flux vectors are scaled by 300 kg m−1 s−1. Shading in panels indicates the frequency of occurrence of (a)–(c) AR and (d)–(f) non-AR conditions.

One concern regarding the ZN definition is that it is somewhat ad hoc, since there is no precise justification for its form; in fact, ZN determined the threshold value 0.3 because it gave the “best” fit to the total moisture flux field computed from the data for one day, 12 October 1991. Changing the threshold parameter gives quite different results: if it is reduced (0.1), then almost all moisture flux worldwide is categorized as “AR” flux and in the northeast Atlantic, AR conditions occur more than 75% of the time; whereas if it is increased (0.5), then the frequency of AR events is so reduced that the AR composite explains only about half of the total flux in the North Pacific. Additionally, the ZN definition does not differentiate between transient and steady moisture transport. The mere fact that moisture transport is much stronger over the oceans than over land, as is the case for transport by the mean circulation in the extratropics (cf. Fig. 1b), is enough to cause many regions to nearly continuously reach the AR threshold, most obviously in the North Atlantic. In fact, all the regions in Fig. 9a where the AR conditions occur at least 20% of the time are also regions where the transport by the mean alone passes ZN ’s AR test. This sensitivity to an arbitrary parameter complicates any diagnosis of how ARs contribute to the total moisture transport.

Note from Fig. 7 that the moisture flux is typically strongest in regions where moisture anomalies are large. This relationship between moisture and moisture flux is fairly general in the extratropics: poleward of about 30°, the correlation between and |Q| is ~0.7–0.85 during DJF (DJF/JJA) in the Northern (Southern) Hemisphere and ~0.5–0.75 during JJA in the Northern Hemisphere (not shown) over the entire 4-times-daily dataset. This correlation supports the approach taken in past AR studies to use IWV as a proxy for Q (Ralph et al. 2004, 2005, 2011; N08; Neiman et al. 2008a). In these studies, which were focused on determining ARs in midlatitudes (often using SSM/I data), ARs were defined using both a threshold value for of 2 cm and a key spatial pattern requirement that retained only narrow plumes that were >2000 km long and <1000 km wide. We defined a somewhat similar criterion by adapting the ZN approach to , specifying that rivers exist wherever ≥ 2 cm and additionally that mean + 0.3(maxmean) on a latitude circle but with no other shape criterion, and then applied this definition to the entire dataset to produce the results shown in Figs. 9b and 9e. Clearly, this AR definition is useful for some areas, including the West Coast of the United States, but it is not general enough for worldwide extratropical application, particularly at much higher latitudes in the Pacific and Atlantic Oceans, where typical values of are less than 2 cm yet strong winds still yield pronounced moisture fluxes. In contrast, it is interesting to note that all wintertime landfalling AR events that N08 identified from their 8-yr dataset are associated with large moisture and flux LF anomalies reaching the coast (e.g., left panels of Fig. 8).

An important aspect of the above-mentioned criteria is that the AR region is defined as relatively narrow, which introduces an element of subjectivity; namely, how narrow is narrow enough? Also, as noted by Bao et al. (2006), ARs are generally coincident with strong surface convergence, so when narrowness is associated with frontal dynamics, it may not be a necessary condition. Nevertheless, these definitions capture the essence of extratropical moisture transport as is seen in Fig. 7, since they identify plumes of moisture with regions of intense poleward moisture transport, as in Ralph et al. (2004). This leads us to categorize AR conditions as the occurrence of episodic poleward-moving moisture plumes, without requiring a shape requirement. Figure 9c shows the results so obtained, by compositing over all times/locations for which the 4-times-daily (unfiltered) and poleward low-level meridional wind anomalies are both positive, with defined as the meridional wind vertically integrated between the surface and σ = 0.85 (results are insensitive to the integral ’s precise upper bound). It is striking that the extratropical results are quite similar to Figs. 9a and 9b; in fact, the composite poleward moisture flux is even greater than the mean. Composite fluxes in the three remaining groupings (anomalous and < 0, > 0 and < 0, < 0 and > 0) are each weaker and contribute roughly equally to the remaining zonal flux in the extratropics, with a somewhat larger equatorward component from the anomalous and < 0 grouping (not shown).

The composite in Fig. 9c shows that the extratropical moisture transport is associated primarily with the anomalous poleward advection of positive moisture anomalies. At any given time, then, ARs are indeed those regions where most of the extratropical moisture flux is located. It is also interesting that the variances of both moisture and meridional wind synoptic anomalies (shown for wintertime in Fig. 10) lie within the region of strongest climatological meridional moisture gradient, with mean moisture relatively well mixed both to the north and south (see the top panel of Fig. 10). Since the synoptic moisture transport is predominantly meridional, the AR composite suggests a simple “lateral mixing” argument for the moisture flux (illustrated in Fig. 11a): anomalous poleward wind generates a positive moisture anomaly (a “plume”) that transports moisture poleward, but at the same latitude anomalous equatorward wind does not generate a negative moisture anomaly, so it does not contribute to the transport. This is essentially the converse of the argument Pierrehumbert (1998) makes for the transport of dry extratropical air into the subtropics, so we have appropriated his term to describe the process (see also Pierrehumbert 2002; Caballero and Langen 2005; O ’Gorman and Schneider 2008). That is, a simple scale analysis for extratropical moisture transport is QQy ~ υΔ, where Δ is the difference (decrease) in precipitable water from the tropics to the extratropics within the storm-track region (as illustrated in Fig. 11a). From Fig. 10, typical values in the Pacific are ~5 m s−1 and Δ ~ 10 mm, yielding ~ 10 mm and Qy ~ 50 kgm−1s−1, both consistent with observations.

Fig. 10.

Moisture and meridional wind DJF climatology. (top) Mean DJF climate of (shading, contour interval 5 mm), variance of synoptic precipitable water anomalies (black, contour interval = 10 mm2), and variance of LF precipitable water anomalies (white, contour interval = 10 mm2). (bottom) Mean DJF climate of low-level (integrated between σ = 0.85 and 1) wind (shading, contour interval 1 m s−1), variance of synoptic low-level wind anomalies (black, contour interval = 5 m2 s−2), and variance of LF low-level wind anomalies (white, contour interval = 5 m2 s−2).

Fig. 10.

Moisture and meridional wind DJF climatology. (top) Mean DJF climate of (shading, contour interval 5 mm), variance of synoptic precipitable water anomalies (black, contour interval = 10 mm2), and variance of LF precipitable water anomalies (white, contour interval = 10 mm2). (bottom) Mean DJF climate of low-level (integrated between σ = 0.85 and 1) wind (shading, contour interval 1 m s−1), variance of synoptic low-level wind anomalies (black, contour interval = 5 m2 s−2), and variance of LF low-level wind anomalies (white, contour interval = 5 m2 s−2).

Fig. 11.

Schematics of extratropical (a) synoptic transport and (b) LF transport. In (a), “lateral mixing” picture of synoptic moisture transport. At a given latitude indicated by the dashed line, a parcel advected by a poleward wind anomaly υ’ (red arrow) will be coming from the equatorward side, so it will have moisture + Δ, meaning that compared to the surrounding air it will have a moisture anomaly = Δ, and there is poleward moisture flux υ’Δ (green arrow). Conversely, a parcel advected by an equatorward wind anomaly v ’ (red arrow) will be coming from the poleward side, so its moisture is the same as the surrounding air, meaning that = 0, and there is no moisture flux. In (b) changes in the surface zonal winds (red lines) associated with anomalous (left) deepening or (right) weakening of the Aleutian low on LF time scales will drive anomalous surface evaporation (orange oval, indicating moisture source), as meridional wind anomalies will advect anomalously dry air equatorward and moist air poleward. This gives rise to moisture flux (green arrows) that is both northwestward and northeastward from the source region to the sinks (blue circles), since the anomalous moisture gradient is in the same direction as the wind anomaly. Since a LF anomaly of either sign will lead to the same pattern of moisture transport, on average this anomaly should contribute to mean transport.

Fig. 11.

Schematics of extratropical (a) synoptic transport and (b) LF transport. In (a), “lateral mixing” picture of synoptic moisture transport. At a given latitude indicated by the dashed line, a parcel advected by a poleward wind anomaly υ’ (red arrow) will be coming from the equatorward side, so it will have moisture + Δ, meaning that compared to the surrounding air it will have a moisture anomaly = Δ, and there is poleward moisture flux υ’Δ (green arrow). Conversely, a parcel advected by an equatorward wind anomaly v ’ (red arrow) will be coming from the poleward side, so its moisture is the same as the surrounding air, meaning that = 0, and there is no moisture flux. In (b) changes in the surface zonal winds (red lines) associated with anomalous (left) deepening or (right) weakening of the Aleutian low on LF time scales will drive anomalous surface evaporation (orange oval, indicating moisture source), as meridional wind anomalies will advect anomalously dry air equatorward and moist air poleward. This gives rise to moisture flux (green arrows) that is both northwestward and northeastward from the source region to the sinks (blue circles), since the anomalous moisture gradient is in the same direction as the wind anomaly. Since a LF anomaly of either sign will lead to the same pattern of moisture transport, on average this anomaly should contribute to mean transport.

This picture can be somewhat extended in two ways. First, note that the maximum in amplitude should coincide with the zero line of moisture flux divergence: increases (decreases) as air south (north) of the maximum moves poleward, so that Qy must also increase (decrease) with latitude, resulting in moisture divergence (convergence) and hence a moisture source (sink). This is consistent with the synoptic advection term in Fig. 5; also, enhanced evaporation is proportional to surface wind anomalies (e.g., Alexander and Scott 1997). Second, as the moisture plumes are advected poleward, they are simultaneously advected by the time-mean flow (as in Fig. 7) so that they do not represent strictly meridional transport. Thus, the AR composite does not simply resemble the synoptic transport but also contains much of the transport by the mean. It follows that when a moisture plume is not present, advection due to the strong zonal winds should be much less. Note that this simple picture is still incomplete, since it does not include the synoptic moisture sink due to the divergence term (Fig. 5), that is, because of low-level convergence and rising motion (e.g., into the moist conveyor belt; Bao et al. 2006) leading to precipitation.

5. LF transport over the wintertime extratropical oceans

The lateral mixing argument of the previous section does not appear to be consistent with the observed variability of LF moisture and circulation anomalies, also shown in Fig. 10, or their associated mean moisture transport. Most of the extratropical meridional wind variability on LF time scales, associated with changes in regional zonal jets and storm tracks, is located in the northeastern portions of the Atlantic and Pacific basins, away from regions of strong mean moisture gradient. Yet as seen in Fig. 5, LF moisture flux divergence is predominantly associated with the advection term. This suggests that a process connected to the typical large-scale LF anomalies of both the Pacific and Atlantic basins must be driving the extratropical LF transport. For example, over the Pacific a common LF anomaly involves a strengthening or weakening of the Aleutian low with corresponding wind anomalies, as illustrated in Fig. 11b. Changes in the surface zonal winds (red lines) will also change surface evaporation (Cayan 1992; Alexander and Scott 1997), and meridional wind anomalies will advect dry air anomalies equatorward and moist air anomalies poleward. This gives rise to LF moisture flux that is both northwestward and northeastward from the source region, with the anomalous moisture gradient in the same direction as the wind anomaly as in Fig. 11b. Note that a LF anomaly of either sign will lead to the same pattern of moisture transport, so on average this anomaly will contribute to mean transport.

To see if this effect exists in nature, we first computed the principal components (PCs) of (LF zonal wind anomaly integrated from the surface to σ = 0.85) during winter over either the Pacific (120°E–120°W) or the Atlantic (90°W–0°) in the Northern Hemisphere. The global fields of , , and are then regressed onto the leading Pacific PC (PC1/PAC; Fig. 12a) and onto the leading Atlantic PC (PC1/ATL; Fig. 12c). In both basins, anomalous surface westerlies (easterlies) are indeed associated with an anomalous positive (negative) zonal gradient of moisture LF anomalies. For example, from Fig. 12a, at about 30°N, 150°W the regressed wind speed is ~0.8 m s−1 and the moisture gradient is ~.0025 mm km−1, so that the flux convergence is ~.17 mm day−1. This estimate does not take into account the vertical profiles of winds or moisture, so it probably underestimates the total flux. Alternatively, in Fig. 12b we show the composite of and its associated divergence for all high-amplitude PC1/PAC events of both signs, that is, when the amplitude of PC1/PAC was either greater than +2 standard deviations or less than −2 standard deviations. The result is again consistent with the simpler argument outlined in Fig. 11b. In particular, individual composites based on +2σ PC1/PAC and −2σ PC1/PAC (not shown) both have similar, statistically significant patterns of flux and flux divergence, with the same sign, except that the convergence near 165°E is only associated with an anticyclonic anomaly. The corresponding composite for the PC1/ATL is shown in Fig. 12d. In this case, although the individual composites again have similar patterns, the +2σ composite amplitude is much larger.

Fig. 12.

Regression of , (indicated together by wind vectors, in m s−1) and (indicated by color shading) on leading wintertime PC (PC1) defined in (a) the Pacific sector and (c) the Atlantic sector; the sign of PC1 in each sector is chosen so that positive corresponds to a maximum westerly wind anomaly (at about 30°N in the Pacific and 55°N in the Atlantic). Composite of LF moisture transport (vectors) and associated flux divergence (shaded) averaged over both +2 and −2 standard deviation values of PC1 for (b) the Pacific sector and (d) the Atlantic sector; only values that are locally 95% significant (based on 1000 Monte Carlo simulations) are shown.

Fig. 12.

Regression of , (indicated together by wind vectors, in m s−1) and (indicated by color shading) on leading wintertime PC (PC1) defined in (a) the Pacific sector and (c) the Atlantic sector; the sign of PC1 in each sector is chosen so that positive corresponds to a maximum westerly wind anomaly (at about 30°N in the Pacific and 55°N in the Atlantic). Composite of LF moisture transport (vectors) and associated flux divergence (shaded) averaged over both +2 and −2 standard deviation values of PC1 for (b) the Pacific sector and (d) the Atlantic sector; only values that are locally 95% significant (based on 1000 Monte Carlo simulations) are shown.

Previous studies (Cayan 1992; Alexander and Scott 1997) have identified the northeast Pacific and Atlantic as regions of strong latent heat flux exchange between the ocean and atmosphere on LF time scales, and this flux is strong enough to drive much of the North Pacific SST anomaly associated with ENSO (Alexander et al. 2002). The net energy flux associated with the North Pacific moisture source in Fig. 1c is L(EP), which for L = 2.5 × 106 J kg−1 and E − P ~1.7 mm day−1 is ~50 W m−2. In comparison, the total 1968–2006 DJF mean latent heat flux in this location (35°N, 142°W) from the objectively analyzed air–sea fluxes (OAFlux) dataset (Yu et al. 2008) is about 85 W m−2. It is also interesting to note that a local ridge in the DJF mean latent heat flux field from the OAflux (not shown) is centered along a line that extends from about 22°N, 147°W to 46°N, 130°W, which is coincident with the maximum in .

6. Concluding remarks

Although there have been many previous analyses of the atmospheric moisture budget, including those that have demonstrated the importance of transient eddies to meridional moisture transport, it has not been previously shown how synoptic-versus-LF time scales impact climatological moisture transport. An analysis of the seasonal cycle of the mean vertically integrated atmospheric moisture budget using 40 yr of NCEP–NCAR reanalysis data reveals that during the cool season in the extratropics of both hemispheres, LF and synoptic anomalies play a significant role in the atmospheric transport of moisture from ocean to land. This occurs despite the fact that transport by the mean circulation generally has a much larger amplitude because much of the transport by the mean does not move moisture onto land so much as move moisture zonally from the western to the eastern margins of the ocean basins. In some regions, such as the North American southwest, Europe, and Australia, the LF transport is the largest contributor to net wintertime and even annual mean atmospheric moisture. The LF transport is also critical to the Arctic moisture budget throughout the year and reaches maximum amplitude during summer, associated with moisture transport from land to ocean, especially over Eurasia.

In addition, the differences between the LF and synoptic transport patterns are so striking as to suggest that LF transport does not merely represent a red-noise residual of synoptic variability, but that the dynamical processes driving LF transport are fundamentally different from those driving synoptic transport. Despite its relatively small impact in many regions of the globe, LF transport is a key moisture source for continental precipitation during winter. Note that while the sources associated with synoptic transport are fairly similar to the dominant global mean moisture sources, LF transport sources are not and, in fact, in many areas oppose the mean. This suggests that it may be of interest to consider these regions as starting points for Lagrangian analyses, especially for case studies of moisture source regions connected to LF variability.

This paper has also examined the potential role of atmospheric rivers in the global water budget and explored a method to diagnose systematically AR contributions to moisture transport without necessarily including a dependence on width and length (e.g., large values of IWV in long and narrow regions in the extratropics) used in recent diagnostic studies by Ralph et al. (2004, 2005, 2006, 2011) and Neiman et al. (2008a,b). The results verify that ARs are the primary regions where extratropical atmospheric moisture transport occurs. An individual AR event is the sum of its mean, synoptic, and LF components. AR moisture transport over the northern midlatitude oceans then essentially consists of poleward and eastward advection of a moisture plume originating within subtropical source regions, plus additional moisture extracted from the ocean in the western storm-track region by synoptic-scale meridional winds, plus moisture extracted in the northeast part of the basin depending on the state of the LF anomaly (e.g., an intensification of the Aleutian low), and minus the water precipitated out poleward of the storm track across the ocean. In this view, ARs do not simply represent trajectories of moisture transport from the tropics/subtropics, since on average ARs also pick up additional moisture as they cross the oceans, a point similar to the one made in case studies of some AR events by Bao et al. (2006).

It has been shown that on at least a few occasions, some moisture may be transported directly from the tropics within ARs (Bao et al. 2006; Stohl et al. 2008; Ralph et al. 2011). In fact, a moisture source for North America due to LF variability exists in the eastern tropical Pacific, although it may be more relevant for Mexico than regions farther north. But we also find a mechanism with perhaps greater impact that represents at most only an indirect effect of the tropics on extratropical moisture and its transport. During wintertime, ENSO is well known to cool (warm) the North Pacific sea surface by intensifying (weakening) the Aleutian low leading to enhanced (weakened) latent heat flux (the “atmospheric bridge”; Alexander et al. 2002). Our results suggest that during this process, moisture is extracted from the sea surface for LF anomalies of either sign, with much of this moisture transported toward western North America. That is, tropical forcing can produce circulation anomalies that transport additional moisture from an extratropical source while not actually transporting moisture all the way from the tropics. Note that extratropical LF variability of this type can also occur without tropical forcing (e.g., Winkler et al. 2001), and that tropical forcing details may influence the extratropical LF anomalies (e.g., Winkler et al. 2001; Di Lorenzo et al. 2010) and how they interact with the sea surface. Whether this process is likewise important to moisture transported by individual synoptic and LF events, including those ARs that give the appearance of direct transport of moisture from the tropics, is the subject of our current research.

Finally, we note that changes in the hydrological cycle are also fundamental to anthropogenic climate change scenarios, impacting precipitation patterns (e.g., Trenberth 2011), large-scale circulation (Held and Soden 2006), and driving much of the global “warming” itself (e.g., Solomon et al. 2007; Compo and Sardeshmukh 2009). For example, it has been suggested that in a global sense, specific humidity may increase with a Clausius–Clapeyron relationship to increasing surface temperatures but precipitation may not, so that atmospheric circulation must weaken to compensate (Held and Soden 2006). However, where LF and/or synoptic transports oppose transport by the mean, as is the case for some areas in the extratropics and even some tropical convergence zones, it is plausible that increased variability could also weaken precipitation. Also, anthropogenic impacts on the northeast Pacific and Atlantic basins that are key moisture sources for extratropical landmasses, and on the LF variability that drives the moisture transport into land, are uncertain. Thus, understanding potential anthropogenic changes in the earth ’s hydrological cycle likely requires understanding corresponding changes in atmospheric variability, especially on low-frequency time scales.

Acknowledgments

The authors wish to thank Jim Adams for drafting Fig. 11; Chris Winkler for his assistance in generating Fig. A1; Gary Wick for his assistance with the SSM/I data; and Alan Betts, Paul Dirmeyer, and Tapio Schneider for their useful comments.

APPENDIX

Dynamically Consistent Estimates of Wind: The Chi Problem

There are many potential sources of error in computing (1) from analyzed datasets, but perhaps the most important is the notable difference in the divergent wind field between reanalyses. This is particularly true in the tropics (Newman et al. 2000), but it can even be true in the extratropics, such as for the low-level jet (LLJ), which transports a significant fraction of moisture during summer (e.g., Helfand and Schubert 1995). Thus, the most important correction to Q can be made by improving the wind analysis [Trenberth and Guillemot 1995; Wang and Paegle 1996; Min and Schubert 1997; Mo and Higgins 1996; see also Bengtsson et al. (2004), who find only a minor impact of assimilation of humidity observations on the 40-yr European Centre for Medium-Range Weather Forecasts Re-Analysis (ERA-40) hydrological cycle]. Specific humidity corrections on the order of 3% have also been made, but these have typically been applied only to the mean field (Large and Yeager 2009). For daily averaged values, we find that for the years 1997–2007 over the oceans, from the reanalysis and IWV from SSM/I are generally well correlated (0.8–0.95) throughout the extratropics and less well (0.5–0.7) in the tropics, although since tropical moisture variance is much less, it might be more susceptible to small errors (cf. Fig. 10a).

How can we better estimate the wind field? One approach is to note that the error in the analyzed wind fields is predominantly in the divergent component of the wind and less so for the rotational component, consistent with the fact that the large-scale vorticity analyses produced at different data centers are in much better agreement than the corresponding divergence analyses (e.g., Newman et al. 2000). One way to correct the analyzed divergence is by constraining the winds to minimize imbalances in both the mass and vorticity budgets, thus enforcing dynamical consistency on the divergent circulation. This approach, known as the “chi problem” (Sardeshmukh 1993), has been successfully used to correct tropical divergence fields (Sardeshmukh and Liebmann 1993; Sardeshmukh et al. 1999), but the approach is applicable globally. A long-term global heating dataset developed using the “chi corrected” horizontal wind and vertical velocity fields, where heating is then estimated as a residual in the heat budget, has been used for studies of short-term climate variability in and related to the tropics (Winkler et al. 2001; Lin et al. 2004; Newman and Sardeshmukh 2008; Newman et al. 2009), since we have found that this technique yields improved diabatic heating estimates (Sardeshmukh et al. 1999). It can similarly improve moisture flux estimates. Fig. A1a shows that in the tropics, the chi-corrected vertical profile of the November–February 1992/93 mean “moisture sink” (Q2; Yanai et al. 1973) compares better with Tropical Ocean and Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE) observations (Johnson and Ciesielski 2000) than does the same quantity computed from the NCEP reanalysis. The chi correction also acts to slightly lower the low-level jet altitude over the central United States, as shown in Fig. A1b for July 1993. A similar change is evident during June–August 1994 (not shown), which is consistent with profiler data showing the jet centered about 50 mb lower than in the reanalysis data (Higgins et al. 1997). The chi correction is a conservative adjustment to wind, well within observational error. Yet, during the warm season, a slightly stronger and lower chi-corrected low-level jet results in a larger estimate of PE in the Great Plains (Fig. A1c), consistent with earlier work suggesting a large dry bias in the reanalysis in the North American warm season (e.g., Mo and Higgins 1996; Yeh et al. 1998; Roads and Betts 2000; Lenters et al. 2000).

Fig. A1.

Comparison of chi-corrected and NCEP moisture flux divergences. (a) Vertical profile of seasonally averaged (1 Nov 1992–28 Feb 1993) moisture sink Q2 from observations (black line), chi-corrected fields (red line) and NCEP reanalysis (blue line). Units are K day−1. Observations are taken from Johnson and Ciesielski (2000) over the intensive flux array (IFA) of COARE. NCEP and chi-corrected fields are measured at an analysis grid point (1.4°S, 155°E) near the center of the IFA region. (b) Latitude-sigma cross sections of July 1993 monthly-mean 1200 UTC meridional wind speed, averaged between 95.5° and 98.5°W. (left) Chi-corrected total meridional wind. Contour interval is 1.2 m s−1; red shading indicates positive values and starts at 1.2 m s−1. (right) Difference fields (chi-corrected winds minus NCEP reanalysis winds). Contour interval is 0.4 m s−1; red shading indicates positive values and blue shading indicates negative values. Shading starts at +0.4 m s−1. Zero contour has been omitted for clarity. (c) Seasonal cycle of PE, area averaged over a region similar to the Mississippi basin region defined by Roads et al. (1994).

Fig. A1.

Comparison of chi-corrected and NCEP moisture flux divergences. (a) Vertical profile of seasonally averaged (1 Nov 1992–28 Feb 1993) moisture sink Q2 from observations (black line), chi-corrected fields (red line) and NCEP reanalysis (blue line). Units are K day−1. Observations are taken from Johnson and Ciesielski (2000) over the intensive flux array (IFA) of COARE. NCEP and chi-corrected fields are measured at an analysis grid point (1.4°S, 155°E) near the center of the IFA region. (b) Latitude-sigma cross sections of July 1993 monthly-mean 1200 UTC meridional wind speed, averaged between 95.5° and 98.5°W. (left) Chi-corrected total meridional wind. Contour interval is 1.2 m s−1; red shading indicates positive values and starts at 1.2 m s−1. (right) Difference fields (chi-corrected winds minus NCEP reanalysis winds). Contour interval is 0.4 m s−1; red shading indicates positive values and blue shading indicates negative values. Shading starts at +0.4 m s−1. Zero contour has been omitted for clarity. (c) Seasonal cycle of PE, area averaged over a region similar to the Mississippi basin region defined by Roads et al. (1994).

REFERENCES

REFERENCES
Alexander
,
M. A.
, and
J. D.
Scott
,
1997
:
Surface flux variability over the North Pacific and North Atlantic Oceans
.
J. Climate
,
10
,
2963
2978
.
Alexander
,
M. A.
,
I.
Bladé
,
M.
Newman
,
J. R.
Lanzante
,
N.-C.
Lau
, and
J. D.
Scott
,
2002
:
The atmospheric bridge: The influence of ENSO teleconnections on air–sea interaction over the global oceans
.
J. Climate
,
15
,
2205
2231
.
Bao
,
J.-W.
,
S. A.
Michelson
,
P. J.
Neiman
,
F. M.
Ralph
, and
J. M.
Wilczak
,
2006
:
Interpretation of enhanced integrated water vapor bands associated with extratropical cyclones: Their formation and connection to tropical moisture
.
Mon. Wea. Rev.
,
134
,
1063
1080
.
Bengtsson
,
L.
,
K. I.
Hodges
, and
S.
Hagemann
,
2004
:
Sensitivity of large-scale atmospheric analyses to humidity observations and its impact on the global water cycle and tropical and extratropical weather systems in ERA40
.
Tellus
,
56A
,
202
217
.
Betts
,
A. K.
,
J. H.
Ball
, and
P.
Viterbo
,
1999
:
Basin-scale surface water and energy budgets for the Mississippi from the ECMWF reanalysis
.
J. Geophys. Res.
,
104
(
D16
),
19 293
19 306
.
Blackmon
,
M. L.
,
J. M.
Wallace
,
N.-C.
Lau
, and
S. L.
Mullen
,
1977
:
An observational study of the Northern Hemisphere wintertime circulation
.
J. Atmos. Sci.
,
34
,
1040
1053
.
Borges
,
M. D.
, and
P. D.
Sardeshmukh
,
1995
:
Barotropic Rossby wave dynamics of zonally varying upper-level flows during northern winter
.
J. Atmos. Sci.
,
52
,
3779
3796
.
Caballero
,
R.
, and
P. L.
Langen
,
2005
:
The dynamic range of poleward energy transport in an atmospheric general circulation model
.
Geophys. Res. Lett.
,
32
,
L02705
,
doi:10.1029/2004GL021581
.
Cayan
,
D. R.
,
1992
:
Latent and sensible heat flux anomalies over the northern oceans: The connection to monthly atmospheric circulation
.
J. Climate
,
5
,
354
369
.
Chang
,
E. K. M.
,
S.
Lee
, and
K. L.
Swanson
,
2002
:
Storm track dynamics
.
J. Climate
,
15
,
2163
2183
.
Compo
,
G. P.
, and
P. D.
Sardeshmukh
,
2009
:
Oceanic influences on recent continental warming
.
Climate Dyn.
,
32
,
333
342
,
doi:10.1007/s00382-008-0448-9
.
Delcroix
,
T.
,
C.
Henin
,
V.
Porte
, and
P.
Arkin
,
1996
:
Precipitation and sea-surface salinity in the tropical Pacific Ocean
.
Deep-Sea Res. I
,
43
,
1123
1141
,
doi:10.1016/0967-0637(96)00048-9
.
Di Lorenzo
,
E.
,
K. M.
Cobb
,
J. C.
Furtado
,
N.
Schneider
,
B. T.
Anderson
,
A.
Bracco
,
M. A.
Alexander
, and
D. J.
Vimont
,
2010
:
Central Pacific El Niño and decadal climate change in the North Pacific Ocean
.
Nat. Geosci.
,
3
,
762
765
,
doi:10.1038/NGEO984
.
Dirmeyer
,
P. A.
, and
K. L.
Brubaker
,
2007
:
Characterization of the global hydrologic cycle from a back-trajectory analysis of atmospheric water vapor
.
J. Hydrometeor.
,
8
,
20
37
.
Dole
,
R.
, and
Coauthors
,
2011
:
Was there a basis for anticipating the 2010 Russian heat wave?
Geophys. Res. Lett.
,
38
,
L06702
,
doi:10.1029/2010GL046582
.
Drumond
,
A.
,
R.
Nieto
, and
L.
Gimeno
,
2011
:
On the contribution of the tropical Western Hemisphere warm pool source of moisture to the Northern Hemisphere precipitation through a Lagrangian approach
.
J. Geophys. Res.
,
116
,
D00Q04
,
doi:10.1029/2010JD015397
.
Eckhardt
,
S.
,
A.
Stohl
,
H.
Wernli
,
P.
James
,
C.
Forster
, and
N.
Spichtinger
,
2004
:
A 15-year climatology of warm conveyor belts
.
J. Climate
,
17
,
218
237
.
Gimeno
,
L.
,
A.
Drummond
,
R.
Nieto
,
R. M.
Trigo
, and
A.
Stohl
,
2010
:
On the origin of continental precipitation
.
Geophys. Res. Lett.
,
37
,
L13804
,
doi:10.1029/2010GL043712
.
Gutowski
,
W. J.
,
Y.
Chen
, and
Z.
Otles
,
1997
:
Atmospheric water vapor transport in NCEP–NCAR reanalyses: Comparison with river discharge in the central United States
.
Bull. Amer. Meteor. Soc.
,
78
,
1957
1969
.
Held
,
I. M.
, and
B. J.
Soden
,
2006
:
Robust responses of the hydrological cycle to global warming
.
J. Climate
,
19
,
5686
5699
.
Helfand
,
H. M.
, and
S. D.
Schubert
,
1995
:
Climatology of the simulated Great Plains low-level jet and its contribution to the continental moisture budget of the United States
.
J. Climate
,
8
,
784
806
.
Higgins
,
R. W.
,
K. C.
Mo
, and
S. D.
Schubert
,
1996
:
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
.
Higgins
,
R. W.
,
Y.
Yao
,
E. S.
Yarosh
,
J. E.
Janowiak
, and
K. C.
Mo
,
1997
:
Influence of the Great Plains low-level jet on summertime precipitation and moisture transport over the central United States
.
J. Climate
,
10
,
481
507
.
Huang
,
R. X.
,
1993
:
Real freshwater flux as a natural boundary condition for the salinity balance and thermohaline circulation forced by evaporation and precipitation
.
J. Phys. Oceanogr.
,
23
,
2428
2446
.
Johnson
,
R. H.
, and
P. E.
Ciesielski
,
2000
:
Rainfall and radiative heating rate estimates from TOGA COARE atmospheric budgets
.
J. Atmos. Sci.
,
57
,
1497
1514
.
Knippertz
,
P.
, and
H.
Wernli
,
2010
:
A Lagrangian climatology of tropical moisture exports to the Northern Hemispheric extratropics
.
J. Climate
,
23
,
987
1003
.
Large
,
W. G.
, and
S. G.
Yeager
,
2009
:
The global climatology of an interannually varying air–sea flux data set
.
Climate Dyn.
,
33
,
341
364
.
Lenters
,
J. D.
,
M. T.
Coe
, and
J. A.
Foley
,
2000
:
Surface water balance of the continental United States, 1963–1995: Regional evaluation of a terrestrial biosphere model and the NCEP/NCAR reanalysis
.
J. Geophys. Res.
,
105
(
D17
),
22 393
22 425
.
Lin
,
J.
,
B.
Mapes
,
M.
Zhang
, and
M.
Newman
,
2004
:
Stratiform precipitation, vertical heating profiles, and the Madden–Julian oscillation
.
J. Atmos. Sci.
,
61
,
296
309
.
Min
,
W.
, and
S.
Schubert
,
1997
:
The climate signal in regional moisture fluxes: A comparison of three global data assimilation products
.
J. Climate
,
10
,
2623
2642
.
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
.
Mo
,
K. C.
,
M.
Chelliah
,
M. L.
Carrera
,
R. W.
Higgins
, and
W.
Ebisuzaki
,
2005
:
Atmospheric moisture transport over the United States and Mexico as evaluated in the NCEP regional reanalysis
.
J. Hydrometeor.
,
6
,
710
728
.
Neiman
,
P. J.
,
F. M.
Ralph
,
G. A.
Wick
,
Y.-H.
Kuo
,
T.-K.
Wee
,
Z.
Ma
,
G. H.
Taylor
, and
M. D.
Dettinger
,
2008a
:
Diagnosis of an intense atmospheric river impacting the Pacific Northwest: Storm summary and offshore vertical structure observed with COSMIC satellite retrievals
.
Mon. Wea. Rev.
,
136
,
4398
4420
.
Neiman
,
P. J.
,
F. M.
Ralph
,
G. A.
Wick
,
J.
Lundquist
, and
M. D.
Dettinger
,
2008b
:
Meteorological characteristics and overland precipitation impacts of atmospheric rivers affecting the west coast of North America based on eight years of SSM/I satellite observations
.
J. Hydrometeor.
,
9
,
22
47
.
Newman
,
M.
, and
P. D.
Sardeshmukh
,
2008
:
Tropical and stratospheric influences on extratropical short-term climate variability
.
J Climate
,
21
,
4326
4347
.
Newman
,
M.
,
P. D.
Sardeshmukh
, and
J. W.
Bergman
,
2000
:
An assessment of the NCEP, NASA and ECMWF, reanalyses over the tropical west Pacific warm pool
.
Bull. Amer. Meteor. Soc.
,
81
,
41
48
.
Newman
,
M.
,
P. D.
Sardeshmukh
, and
C.
Penland
,
2009
:
How important is air–sea coupling in ENSO and MJO evolution?
J. Climate
,
22
,
2958
2977
.
O ’Gorman
,
P. A.
, and
T.
Schneider
,
2008
:
The hydrological cycle over a wide range of climates simulated with an idealized GCM
.
J. Climate
,
21
,
3815
3832
.
Pauluis
,
O.
,
T.
Shaw
, and
F.
Laliberté
,
2011
:
A statistical generalization of the transformed Eulerian-mean circulation for an arbitrary vertical coordinate system
.
J. Atmos. Sci.
,
68
,
1766
1783
.
Peixoto
,
J. P.
, and
A. H.
Oort
,
1992
:
Physics of Climate. American Institute of Physics, 520 pp
.
Pierrehumbert
,
R. T.
,
1998
:
Lateral mixing as a source of subtropical water vapor
.
Geophys. Res. Lett.
,
25
,
151
154
.
Pierrehumbert
,
R. T.
,
2002
:
The hydrologic cycle in deep-time climate problems
.
Nature
,
419
,
191
198
.
Ralph
,
F. M.
,
P. J.
Neiman
, and
G. A.
Wick
,
2004
:
Satellite and CALJET aircraft observations of atmospheric rivers over the eastern North Pacific Ocean during the winter of 1997/98
.
Mon. Wea. Rev.
,
132
,
1721
1745
.
Ralph
,
F. M.
,
P. J.
Neiman
, and
R.
Rotunno
,
2005
:
Dropsonde observations in low-level jets over the northeastern Pacific Ocean from CALJET-1998 and PACJET-2001: Mean vertical-profile and atmospheric-river characteristics
.
Mon. Wea. Rev.
,
133
,
889
910
.
Ralph
,
F. M.
,
P. J.
Neiman
,
G. A.
Wick
,
S. I.
Gutman
,
M. D.
Dettinger
,
D. R.
Cayan
, and
A. B.
White
,
2006
:
Flooding on California ’s Russian River: Role of atmospheric rivers
.
Geophys. Res. Lett.
,
33
,
L13801, doi:10.1029/2006GL026689.
Ralph
,
F. M.
,
P. J.
Neiman
,
G. N.
Kiladis
,
K.
Weickman
, and
D. W.
Reynolds
,
2011
:
A multiscale observational case study of a Pacific atmospheric river exhibiting tropical–extratropical connections and a mesoscale frontal wave
.
Mon. Wea. Rev.
,
139
,
1169
1189
.
Roads
,
J. O.
, and
A.
Betts
,
2000
:
NCEP–NCAR and ECMWF reanalysis surface water and energy budgets for the Mississippi River basin
.
J. Hydrometeor.
,
1
,
88
94
.
Roads
,
J. O.
,
S.-C.
Chen
,
A. K.
Guetter
, and
K. P.
Georgakakos
,
1994
:
Large-scale aspects of the United States hydrologic cycle
.
Bull. Amer. Meteor. Soc.
,
75
,
1589
1610
.
Roads
,
J. O.
,
M.
Kanamitsu
, and
R.
Stewart
,
2002
:
CSE water and energy budgets in the NCEP–DOE Reanalysis II
.
J. Hydrometeor.
,
3
,
227
248
.
Sardeshmukh
,
P. D.
,
1993
:
The baroclinic “chi” problem and its application to the diagnosis of atmospheric heating rates
.
J. Atmos. Sci.
,
50
,
1099
1112
.
Sardeshmukh
,
P. D.
, and
B. J.
Hoskins
,
1984
:
Spatial smoothing on the sphere
.
Mon. Wea. Rev.
,
112
,
2524
2529
.
Sardeshmukh
,
P. D.
, and
B. L.
Liebmann
,
1993
:
An assessment of low-frequency variability in the tropics as indicated by some proxies of tropical convection
.
J. Climate
,
6
,
569
575
.
Sardeshmukh
,
P. D.
,
M.
Newman
, and
C. R.
Winkler
,
1999
:
Dynamically consistent estimates of diabatic heating. Proc. 24th Annual Climate Diagnostics and Prediction Workshop, Tucson, AZ, NOAA, 172–175
.
Schneider
,
T.
,
K. L.
Smith
,
P. A.
O ’Gorman
, and
C. C.
Walker
,
2006
:
A climatology of tropospheric zonal-mean water vapor fields and fluxes in isentropic coordinates
.
J. Climate
,
19
,
5918
5933
.
Schneider
,
T.
,
P. A.
O ’Gorman
, and
X. J.
Levine
,
2010
:
Water vapor and the dynamics of climate changes
.
Rev. Geophys.
,
48
,
RG3001
,
doi:10.1029/2009RG000302
.
Serreze
,
M. C.
, and
Coauthors
,
2006
:
The large-scale freshwater cycle of the Arctic
.
J. Geophys. Res.
,
111
,
C11010
,
doi:10.1029/2005JC003424
.
Serreze
,
M. C.
,
A. P.
Barrett
,
A. G.
Slater
,
M.
Steele
,
J.
Zhang
, and
K. E.
Trenberth
,
2007
:
The large-scale energy budget of the Arctic
.
J. Geophys. Res.
,
112
,
D11122
,
doi:10.1029/2006JD008230
.
Shaw
,
T. A.
, and
O.
Pauluis
,
2012
:
Tropical and subtropical meridional latent heat transports by disturbances to the zonal mean and their role in the general circulation
.
J. Atmos. Sci.
,
69
,
1872
1889
.
Simmons
,
A. J.
,
J. M.
Wallace
, and
G. W.
Branstator
,
1983
:
Barotropic wave propagation and instability, and atmospheric teleconnection patterns
.
J. Atmos. Sci.
,
40
,
1363
1392
.
Solomon
,
S.
,
D.
Qin
,
M.
Manning
,
M.
Marquis
,
K.
Averyt
,
M. M. B.
Tignor
,
H. L.
Miller
Jr.
, and
Z.
Chen
, Eds.,
2007
:
Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp
.
Stohl
,
A.
, and
P.
James
,
2005
:
A Lagrangian analysis of the atmospheric branch of the global water cycle. Part II: Moisture transports between earth ’s ocean basins and river catchments
.
J. Hydrometeor.
,
6
,
961
984
.
Stohl
,
A.
,
C.
Forster
, and
H.
Sodemann
,
2008
:
Remote sources of water vapor forming precipitation on the Norwegian west coast at 60°N-A tale of hurricanes and an atmospheric river
.
J. Geophys. Res.
,
113
,
D05102
,
doi:10.1029/2007JD009006
.
Trenberth
,
K. E.
,
1991
:
Climate diagnostics from global analyses: Conservation of mass in ECMWF analyses
.
J. Climate
,
4
,
707
722
.
Trenberth
,
K. E.
,
2011
:
Changes in precipitation with climate change
.
Climate Res.
,
47
,
123
138
,
doi:10.3354/cr00953
.
Trenberth
,
K. E.
, and
C. J.
Guillemot
,
1995
:
Evaluation of the global atmospheric moisture budget as seen from analyses
.
J. Climate
,
8
,
2255
2272
.
Trenberth
,
K. E.
, and
C. J.
Guillemot
,
1998
:
Evaluation of the atmospheric moisture and hydrological cycle in the NCEP/NCAR reanalyses
.
Climate Dyn.
,
14
,
213
231
.
Trenberth
,
K. E.
,
L.
Smith
,
T.
Qian
,
A.
Dai
, and
J. T.
Fasullo
,
2007
:
Estimates of the global water budget and its annual cycle using observational and model data
.
J. Hydrometeor.
,
8
,
758
769
.
Trenberth
,
K. E.
,
J. T.
Fasullo
, and
J.
Mackaro
,
2011
:
Atmospheric moisture transports from ocean to land and global energy flows in reanalyses
.
J. Climate
,
24
,
4907
4924
.
Tyrlis
,
E.
, and
B. J.
Hoskins
,
2008
:
Aspects of a Northern Hemisphere atmospheric blocking climatology
.
J. Atmos. Sci.
,
65
,
1638
1652
.
van Loon
,
H.
,
1967
:
The half-yearly oscillation in middle and high southern latitudes and the coreless winter
.
J. Atmos. Sci.
,
24
,
472
486
.
Wang
,
M.
, and
J.
Paegle
,
1996
:
Impact of analysis uncertainty upon regional atmospheric moisture flux
.
J. Geophys. Res.
,
101
(
D3
),
7291
7303
.
Whitaker
,
J. S.
, and
P. D.
Sardeshmukh
,
1998
:
A linear theory of extratropical synoptic eddy statistics
.
J. Atmos. Sci.
,
55
,
237
258
.
Winkler
,
C. R.
,
M.
Newman
, and
P. D.
Sardeshmukh
,
2001
:
A linear model of wintertime low-frequency variability. Part I: Formulation and forecast skill
.
J. Climate
,
14
,
4474
4494
.
Yanai
,
M. S.
,
S.
Esbensen
, and
J. H.
Chu
,
1973
:
Determination of bulk properties of tropical cloud clusters from large-scale heat and moisture budgets
.
J. Atmos. Sci.
,
30
,
611
627
.
Yeh
,
P. J.-F.
,
M.
Irizarry
, and
E. A. B.
Eltahir
,
1998
:
Hydroclimatology of Illinois: A comparison of monthly evapotranspiration estimates based on atmospheric water balance and soil water balance
.
J. Geophys. Res.
,
103
(
D16
),
19 823
19 837
.
Yu
,
L.
,
X.
Jin
, and
R. A.
Weller
,
2008
:
Multidecade global flux datasets from the Objectively Analyzed Air-sea Fluxes (OAFlux) Project: Latent and sensible heat fluxes, ocean evaporation, and related surface meteorological variables. Woods Hole Oceanographic Institution, OAFlux Project Tech. Rep. OA-2008-01, 64 pp.
Zhu
,
Y.
, and
R. E.
Newell
,
1998
:
A proposed algorithm for moisture fluxes from atmospheric rivers
.
Mon. Wea. Rev.
,
126
,
725
735
.