Abstract

The role of moisture for extratropical atmospheric dynamics is particularly pronounced within warm conveyor belts (WCBs), which are characterized by intense latent heat release and precipitation formation. Based on the WCB climatology for the period 1979–2010 presented in Part I, two important aspects of the WCB moisture cycle are investigated: the evaporative moisture sources and the relevance of WCBs for total and extreme precipitation. The most important WCB moisture source regions are the western North Atlantic and North Pacific in boreal winter and the South Pacific and western South Atlantic in boreal summer. The strongest continental moisture source is South America. During winter, source locations are mostly local and over the ocean, and the associated surface evaporation occurs primarily during 5 days prior to the start of the WCB ascent. Long-range transport and continental moisture recycling are much more important in summer, when a substantial fraction of the evaporation occurs more than 10 days before the ascent. In many extratropical regions, WCB moisture supply is related to anomalously strong surface evaporation, enforced by low relative humidity and high winds over the ocean. WCBs are highly relevant for total and extreme precipitation in many parts of the extratropics. For instance, the percentage of precipitation extremes directly associated with a WCB is higher than 70%–80% over southeastern North America, Japan, and large parts of southern South America. A proper representation of WCBs in weather forecast and climate models is thus essential for the correct prediction of extreme precipitation events.

1. Introduction

Warm conveyor belts (WCBs) are coherent, ascending airstreams associated with extratropical cyclones (Harrold 1973; Carlson 1980; Browning 1986). They typically originate in the moist marine boundary layer of a cyclone’s warm sector and ascend rapidly along the cold front into the upper troposphere, along with a poleward movement of often several thousands of kilometers. WCBs are usually saturated and contribute substantially to the cyclone’s cloud structure and the related surface precipitation (Browning 1986, 1990). They can ventilate huge amounts of water vapor out of the boundary layer (Eckhardt et al. 2004; Boutle et al. 2011) and transport pollutants into the midlatitude and polar upper troposphere (Arnold et al. 1997; Stohl 2001). Furthermore, WCBs influence cyclone dynamics by inducing positive (negative) potential vorticity anomalies in the lower (upper) troposphere (Stoelinga 1996; Wernli and Davies 1997; Wernli 1997; Pomroy and Thorpe 2000; Joos and Wernli 2012; Schemm et al. 2013; Madonna et al. 2014, Part I of this study, hereafter referred to as M14).

WCBs were originally detected on satellite pictures and with the help of isentropic analyses (see again Harrold 1973; Carlson 1980; Browning 1986). Later on, objective identification criteria were introduced based on three-dimensional kinematic trajectory calculations (Wernli and Davies 1997; Wernli 1997). Such criteria were then applied to produce global WCB climatologies (Stohl 2001; Eckhardt et al. 2004). Recently, M14 compiled a new WCB climatology for the period 1979–2010 based on the European Centre for Medium-Range Weather Forecasts (ECMWF) Interim Re-Analysis (ERA-Interim) dataset (Dee et al. 2011). They found WCBs to be most frequent in the subtropical and midlatitude parts of the western ocean basins. The seasonal cycle of WCB occurrence, with more WCBs in winter than in summer, is more pronounced in the Northern Hemisphere compared to the Southern Hemisphere. M14 also determined the evolution of key parameters like pressure, humidity, and potential vorticity (PV) along the WCB trajectories and analyzed in detail the formation of dynamically relevant PV anomalies. In this second part of the climatology, two important aspects of the WCB moisture cycle are investigated: the evaporative sources of WCB humidity and the relevance of WCBs for total and extreme precipitation.

The coupling between atmospheric dynamics and the moisture cycle is essential for the formation of WCBs. Latent heating due to cloud and precipitation formation is the main driver of their strong cross-isentropic ascent (Joos and Wernli 2012). The typical increase in potential temperature due to this diabatic heat release exceeds 20 K (M14). Schemm et al. (2013) used an idealized model to simulate dry and moist baroclinic waves, and they were able to identify WCBs with an ascent exceeding 600 hPa in 2 days only in the model runs including moisture and the corresponding latent heating. Because of this key role of diabatic processes during cloud formation, a sufficient moisture supply is supposed to be crucial for WCB occurrence and strength. For instance, Eckhardt et al. (2004) speculated that the smaller initial moisture content of WCBs over the Indian Ocean leads to a weaker ascent compared to Atlantic or Pacific systems. Wernli (1997) studied the moisture supply of a WCB by looking at specific humidity (q) changes along 3-day backward trajectories calculated from the starting points of the ascent. He found a moderate increase of q, which was interpreted as moistening by surface fluxes. Schäfler et al. (2011) applied a Lagrangian diagnostic based on analysis data for quantifying the moisture sources of a WCB over southwestern Europe in July 2007. They concluded that errors in surface evaporation or moisture transport may have led to an overestimation of the initial WCB humidity content in the analysis dataset compared to the lidar measurements. Such a moist bias may have substantial consequences for the WCB evolution and thus the large-scale flow and associated weather events downstream (e.g., Massacand et al. 2001; Grams et al. 2011). In this study, the same Lagrangian diagnostic as used by Schäfler et al. (2011) is applied to the WCB climatology of M14. In this way, the climatological distribution of evaporative moisture sources of WCBs is determined, and the typical moisture residence time scale is quantified between evaporation and the start of the WCB ascent. In addition, it is investigated if WCBs are associated with anomalously intense surface evaporation, and the influence of surface temperature, relative humidity, and wind speed on such evaporation anomalies is quantified.

The important role of WCBs for precipitation is known from many case studies (e.g., Browning 1986, 1990). Occasionally, WCBs are related to and supplied with moisture from atmospheric rivers (Sodemann and Stohl 2013), which are excursions of humid air from the subtropics into the mid- and high latitudes that are themselves often associated with heavy precipitation (e.g., Ralph et al. 2006; Lavers et al. 2011). To quantify the WCB contribution to precipitation, Wernli and Davies (1997) and Wernli (1997) estimated the precipitation from decreases of q along trajectories. With the same approach, Eckhardt et al. (2004) showed that climatologically WCB trajectories produce much more precipitation than other trajectories starting in the boundary layer. The quantitative precipitation estimates obtained from this method, however, depend crucially on the number of trajectories and thereby on the specific WCB selection criterion (see again Eckhardt et al. 2004). In this study, the role of WCBs for precipitation is investigated with a modified approach in which a humidity decrease along a WCB trajectory is matched with the predicted ERA-Interim surface precipitation at the same location. This allows us to quantify the WCB contribution to total precipitation as well as the relevance of WCBs for precipitation extremes. Such a quantification complements results from other recent studies that investigated the linkage between precipitation and other synoptic-scale circulation features like fronts and cyclones (Catto et al. 2012; Hawcroft et al. 2012; Pfahl and Wernli 2012).

2. Data and methods

a. Reanalysis data and WCB trajectories

This study applies 6-hourly data from the ERA-Interim reanalysis (Dee et al. 2011) of the European Centre for Medium-Range Weather Forecasts for the period 1979–2010, interpolated to a 1° × 1° longitude–latitude grid. In addition to the fields extracted by M14, skin temperature (SKT), 2-m relative humidity (RH, calculated from 2-m temperature and dewpoint temperature), and 10-m wind velocity (V10) from the reanalysis as well as 6-hourly accumulated surface precipitation and surface latent heat flux (LHF) from twice-daily short-term ERA-Interim forecasts are used. Because of a potential model spinup, the first 6 h of the forecasts are neglected, and the data are taken from forecast steps between 6 and 12 as well as 12 and 18 h. For the investigation of the evaporation anomalies in section 3, standard deviations of the 6-hourly data are calculated for each season at every grid point. The precipitation extremes analyzed in section 4 are defined as the 1% strongest 6-hourly precipitation events at each grid point in the entire analysis period 1979–2010. Note that by using ERA-Interim precipitation data, we focus on extreme events affecting areas of at least one 1° × 1° grid box. Pfahl and Wernli (2012) showed that the timing of such extreme precipitation events is reasonably well represented in the ERA-Interim data in the extratropics. However, in the tropics, where convection plays a more important role, the dataset is less reliable.

M14 identified WCBs based on air mass trajectory calculations with the Lagrangian Analysis Tool (LAGRANTO) (Wernli and Davies 1997). All trajectories with an ascent of more than 600 hPa within 2 days occurring in the vicinity of an extratropical cyclone have been classified as WCBs. These trajectories have then been extended 10 days backward and forward to characterize the pre- and postascent phases, respectively. Here, trajectory positions as well as the evolution of specific humidity during the ascent and preascent phase of the WCBs are used.

b. Moisture source diagnostic

To quantify the climatological moisture sources of WCBs, the Lagrangian source diagnostic developed by Sodemann et al. (2008) is employed. Similar Lagrangian methods have been widely used for studying the sources of precipitation [see Gimeno et al. (2012) and references therein]. Here, the diagnostic is applied for the water vapor at the WCB starting points in a similar setup as used by Pfahl and Wernli (2008) and Schäfler et al. (2011). The method evaluates changes in specific humidity along the 10-day preascent backward trajectories. Moisture gains occurring within the extended atmospheric boundary layer (ERA-Interim boundary layer height multiplied by a factor of 1.5) are associated with evaporation from the underlying surface. A weight is assigned to each moisture uptake according to its contribution to the q value at the starting point of the WCB ascent. For moisture uptakes directly prior to the start of the ascent (without moisture loss due to rainout in between), the weight is simply given by the ratio of the moisture increase and the final specific humidity. If q decreases along the trajectory, the weight of previous moisture uptakes is reduced proportionally because, assuming a well-mixed air parcel, part of the humidity from these uptakes is lost due to rainout [for more details, see Sodemann et al. (2008) and Pfahl and Wernli (2008)]. Finally, the seasonal contribution of moisture from each grid point to the WCB humidity is determined by summing up the respective weighted q values from all trajectories. This is done for all WCBs globally and for WCBs starting in specific regions (see Fig. 2 in M14).

In addition to quantitatively assessing the climatological WCB moisture contribution from each location, it is investigated if this moisture supply is associated with anomalous conditions (e.g., if the surface latent heat flux is larger than normal). At each grid point and for each season, the 6-hourly intervals are determined during which the moisture contribution to WCBs from the respective location is larger than a specified threshold Δqupt = 2 g kg−1. Parameters like LHF, SKT, RH, and V10 are then averaged over these intervals and compared to mean values over all other periods without a WCB moisture contribution from this grid point. Normalized anomalies of these parameters are calculated by dividing the difference between the mean values during WCB moisture contribution and all other time instants by the parameter’s standard deviation. The chosen value for Δqupt is fairly large (i.e., a substantial WCB moisture contribution is required from the respective point). Sensitivity tests with lower thresholds of 1 and 0.5 g kg−1 will also be performed.

c. Precipitation attribution

The contribution of WCBs to precipitation is not assessed by directly translating the decrease in specific humidity along the trajectories into surface precipitation amounts, as done in earlier studies (see section 1). Instead, for each WCB an area of influence is defined, and the predicted ERA-Interim surface precipitation within this area is associated with the WCB. In this way, the sensitivity to the WCB ascent criterion is reduced, because all precipitation falling in the vicinity of the WCB is taken into account (instead of only those stemming from trajectories with a certain vertical ascent). The 6-hourly WCB trajectory segments are selected during which the specific humidity decreases by more than a threshold value Δqpr. All grid points surrounding these trajectory segments are then allocated to the WCB’s area of influence. A rather low threshold value of Δqpr = 1 g kg−1 is chosen to cover most of the WCB precipitation. The sensitivity with respect to this choice will be tested by using a higher value of 3 g kg−1.

The percentage of total precipitation associated with a WCB is then determined at each grid point and for the whole analysis period. In the same way, the percentage of precipitation extremes related to a WCB is calculated. These fields are compared to the temporal frequency of a specific grid point being located within a WCB’s area of influence. The deviation from this frequency can be used as a significance measure of the effect of WCBs on precipitation (cf. Pfahl and Wernli 2012). However, a statistical test is not explicitly performed here, since the existence of a linkage between WCBs and precipitation is obvious anyway, and such a test would not add any information. Finally, the combined effect of WCBs and cyclones on precipitation extremes is evaluated by quantifying the contribution of cyclones to extreme precipitation events with the approach of Pfahl and Wernli (2012). The same cyclone data are used as applied for the WCB identification by M14, which have been obtained with the help of a slightly modified version of the scheme of Wernli and Schwierz (2006). Cyclones are defined as areas within closed sea level pressure contours that contain one or several local sea level pressure minima, with the maximum length of the outermost pressure contour restricted to 7500 km.

3. Moisture sources

a. Source distributions

Figure 1 shows the global climatological distribution of WCB moisture sources in December–February (DJF) and June–August (JJA). The most prominent source regions are located over the oceans, in particular over the subtropical Atlantic and Pacific Oceans. Many aspects of the spatial and seasonal source distribution correspond to the climatological distribution of WCB starting points (purple contours in Fig. 1; M14): The total WCB moisture uptake is larger in winter than in summer as WCBs are more frequent during the cold season. This seasonality is most pronounced over the western North Atlantic and weaker, for example, over the Indian Ocean. There is much more WCB moisture uptake over the western than over the eastern part of the Atlantic Ocean, whereas the spatial distribution of both WCB starting points and moisture sources is slightly more zonal over the Pacific. Furthermore, there are many similarities between the spatial and seasonal patterns of WCB moisture uptake and the seasonal-mean surface latent heat flux (Fig. 2). The maxima in evaporation over the warm ocean currents of the western North Atlantic and North Pacific in DJF correspond to the most prominent WCB moisture sources. These regions in the western parts of the Northern Hemisphere ocean basins are particularly favorable for the occurrence of WCBs due to the pronounced baroclinicity and the frequent development of cyclones. In addition, the strong surface evaporation from the warm ocean provides ample humidity. Over the eastern parts of the Atlantic and Pacific Oceans, the cyclone frequency is also high, but the storms are typically located farther north (Wernli and Schwierz 2006), where surface evaporation is weaker and moisture availability is limited, leading to a reduction in WCB occurrence. Most probably, the limited moisture supply is also the reason for the lack of WCBs in the Southern Ocean storm-track region (M14). Also with respect to evapotranspiration from the land surface, there is some correspondence between Figs. 1 and 2. For instance, the greatly reduced land evapotranspiration during winter explains the absence of WCB moisture sources (and WCB starting points; see M14) over most parts of Eurasia. An interesting feature is the small local maximum of WCB moisture uptake in DJF in the Great Lakes region in North America, which might be associated with lake evaporation (cf. Figs. 1a, 2a). The global maximum of WCB moisture supply from land is found over South America in austral summer, corresponding to intense evapotranspiration in the same region. In the cold season, this moisture uptake maximum is shifted equatorward as the LHF in southern South America is much lower than during DJF (see again Fig. 2). In boreal summer, eastern North America, eastern China, India, and Indochina are other important land sources of WCB humidity. In particular, WCBs ascending over the Himalayas (enforced by the Indian monsoonal circulation; see M14) are supplied with moisture from local evapotranspiration along the foothills of the mountains. Other moisture sources for these WCBs are strong evaporation from the Arabian Sea, especially in the region of the Somali Current, and from the Bay of Bengal.

Fig. 1.

Total moisture uptake in g kg−1 m−2 accumulated over all WCB trajectories during (a) DJF and (b) JJA 1979–2010. Note the nonlinear color scale. The purple contour highlights the main starting regions of the WCB ascent (frequency of starting points > 1%; see M14).

Fig. 1.

Total moisture uptake in g kg−1 m−2 accumulated over all WCB trajectories during (a) DJF and (b) JJA 1979–2010. Note the nonlinear color scale. The purple contour highlights the main starting regions of the WCB ascent (frequency of starting points > 1%; see M14).

Fig. 2.

Mean ERA-Interim surface latent heat flux in W m−2 during (a) DJF and (b) JJA 1979–2010. Negative values denote fluxes from the surface into the atmosphere. The purple contour highlights the main starting regions of the WCB ascent (frequency of starting points > 1%; see M14).

Fig. 2.

Mean ERA-Interim surface latent heat flux in W m−2 during (a) DJF and (b) JJA 1979–2010. Negative values denote fluxes from the surface into the atmosphere. The purple contour highlights the main starting regions of the WCB ascent (frequency of starting points > 1%; see M14).

With the 10-day backward trajectories and the Lagrangian diagnostic applied here, it is not possible to determine the sources of all the WCB humidity. The source regions in Fig. 1 correspond to 71% of the moisture at the WCB starting points in DJF and to 65% in JJA. The remaining humidity has either been contained in the air parcels already 10 days before the WCB ascent or is taken up above the boundary layer (cf. section 2b). The latter cannot directly be associated with surface evaporation and is thus not taken into account in the source diagnostic.

To provide a more detailed regional perspective on the WCB humidity supply, Fig. 3 shows the moisture source regions for WCBs starting in the North Pacific region only (0°–90°N, 100°E–120°W). In contrast to Fig. 1, the color shading here indicates relative contributions to the final WCB humidity (such that the integral over the whole region is normalized to 1). In boreal winter, WCB starting points (purple contours) and the corresponding moisture sources are located only over the ocean. As outlined above, this is most probably due to the lack of moisture supply from the land. A projection of the uptake locations to a relative coordinate system with the starting point of the WCB ascent in the center (Fig. 3c) shows that much of the moisture evaporates in the surroundings of the starting point and that the humidity is mainly transported in the zonal direction. The weighted average distance from the source location to the WCB ascent in DJF is 1200 km. The distribution of uptake times (Fig. 3e) has a pronounced peak about 36 h prior to the start of the ascent, and 88% of the explained moisture evaporates in the 5 days before the ascent. For the North Pacific in DJF, this explained moisture of which sources can be determined with our method amounts to 84% of the total moisture at the WCB starting points. On the contrary, during summer both WCB starting locations and moisture sources are shifted westward (Fig. 3b), indicating substantial humidity contributions from land areas in eastern China and on the Indochina Peninsula. The moisture transport has a stronger meridional component than in DJF (Fig. 3d), and long-distance transport is more important, with an average distance between sources and WCB ascent of 2300 km. This is also evident from the distribution of uptake times (Fig. 3f), which is much flatter than for winter. Only 58% of the explained moisture evaporates within the last 5 days. Significant moisture uptakes still occur 10 days prior to the start of the ascent, and sources can be attributed to only 43% of the total moisture. If also moisture uptakes above the boundary layer are taken into account, this explained fraction increases to 61%, with the spatial and temporal uptake patterns almost unchanged (not shown). This shows that more than one-third of the moisture at the WCB starting points in JJA has already been present in the atmosphere for more than 10 days. Most of the general results shown here for the North Pacific are also valid for WCBs starting in other regions (not shown). During summer, moisture sources are generally more continental, and meridional transport from the tropics and subtropics is more important compared to winter. Typical transport time scales and distances are much larger in summer.

Fig. 3.

Moisture sources of WCBs starting in the North Pacific (0°–90°N, 100°E–120°W). Colors show the relative contribution to total WCB moisture in % (10 000 km)−2 on a (a),(b) geographical grid and (c),(d) relative grid where (0, 0) is the starting location of the WCB ascent. The purple contour in (a),(b) highlights the main WCB starting regions (frequency of starting points > 1%; see M14). Histograms in (e),(f) show the distributions of uptake times (in h) relative to the start of the WCB ascent. Data are for (left) DJF and (right) JJA 1979–2010.

Fig. 3.

Moisture sources of WCBs starting in the North Pacific (0°–90°N, 100°E–120°W). Colors show the relative contribution to total WCB moisture in % (10 000 km)−2 on a (a),(b) geographical grid and (c),(d) relative grid where (0, 0) is the starting location of the WCB ascent. The purple contour in (a),(b) highlights the main WCB starting regions (frequency of starting points > 1%; see M14). Histograms in (e),(f) show the distributions of uptake times (in h) relative to the start of the WCB ascent. Data are for (left) DJF and (right) JJA 1979–2010.

b. Anomaly conditions

Figure 4 shows the anomalous surface latent heat flux during humidity supply for a WCB from the respective grid point. To focus on grid points with a robust statistic, the field is only shown in the major WCB moisture source regions where during more than 1% of the 6-hourly intervals in the respective season a WCB moisture uptake of more than Δqupt = 2 g kg−1 occurs. The LHF anomaly is mostly negative, indicating that more water evaporates from the surface during WCB moisture uptake compared to other periods. The strongest anomalies of more than 100 W m−2 are found over land, in particular over eastern Asia in JJA and South America in both seasons. Over the oceans, there is a meridional gradient with stronger anomalies farther away from the equator. No substantial LHF anomalies are found in the tropics (equatorward of 20°), around Australia in JJA and over the East China Sea in DJF. Most of these regions without large LHF anomalies are associated with strong seasonal-mean evaporation (see Fig. 2). For comparing the magnitude of the anomalies with the overall variability of the LHF, they are normalized with the 6-hourly standard deviations (see section 2). Figures 5a and 5b show these normalized LHF anomalies for DJF and JJA, respectively. Anomalies of more than 1 standard deviation (sd) are found over land and over parts of the ocean in summer. During winter, oceanic anomalies are mostly between 0.5 and 1 sd. It should be noted here that also for these more moderate anomalies between 0.5 and 1 sd, the difference in mean LHF between WCB moisture uptakes and all other periods is statistically highly significant (based on a two-sided t test and a 99% confidence interval). A comparison of Figs. 4, 5a, 5b, and 2 shows that the largest oceanic LHF anomalies are found in regions where the seasonal-mean evaporation is relatively weak.

Fig. 4.

Difference in LHF (W m−2) between 6-hourly intervals with substantial WCB moisture contribution from the respective grid point [uptake > 2 g kg−1 (6 h)−1] and all other time instants during (a) DJF and (b) JJA 1979–2010. Negative values denote stronger surface evaporation during WCB moisture uptake than during other times. Grid points where the frequency of WCB moisture uptake is <1% are masked in gray.

Fig. 4.

Difference in LHF (W m−2) between 6-hourly intervals with substantial WCB moisture contribution from the respective grid point [uptake > 2 g kg−1 (6 h)−1] and all other time instants during (a) DJF and (b) JJA 1979–2010. Negative values denote stronger surface evaporation during WCB moisture uptake than during other times. Grid points where the frequency of WCB moisture uptake is <1% are masked in gray.

Fig. 5.

Difference plots as in Fig. 4, but with all variables normalized with their 6-hourly standard deviations for (left) DJF and (right) JJA. Variables shown are (a),(b) LHF, (c),(d) SKT, (e),(f) RH, and (g),(h) V10.

Fig. 5.

Difference plots as in Fig. 4, but with all variables normalized with their 6-hourly standard deviations for (left) DJF and (right) JJA. Variables shown are (a),(b) LHF, (c),(d) SKT, (e),(f) RH, and (g),(h) V10.

Altogether, this indicates that on the one hand there are areas like the tropics and the East China Sea in boreal winter where the mean surface evaporation is strong enough to supply the moisture for a WCB independent of the actual synoptic conditions. On the other hand, over land and in regions with weaker mean LHF, the supply of moisture for a WCB is associated with a temporal anomaly of surface evaporation and thus with specific synoptic conditions. According to simple bulk formulas (see, e.g., Liu et al. 1979), the LHF can be expressed in terms of near-surface wind velocity (which relates to the turbulence conditions) and the specific humidity gradient between the surface and the atmosphere. Over the ocean, the latter mainly depends on skin temperature and near-surface relative humidity. Over land, soil moisture availability plays an additional role, which is not investigated further here. Figures 5c–h show the anomalies of SKT, RH, and V10 associated with WCB moisture uptake. Positive SKT anomalies are mainly found over land (Figs. 5c,d) together with low near-surface RH (Figs. 5e,f), both favoring enhanced evaporation (assuming sufficient soil moisture supply). Note, however, that the large continental SKT, RH, and LHF anomalies are probably related to the strong daily cycle of these variables, suggesting that WCB moisture uptakes over land occur predominantly during daytime. The lack of positive SKT anomalies over the ocean indicates that the enhanced evaporation during WCB moisture uptake is not related to a warmer sea surface. On the contrary, in the western North Pacific there is an area of anomalously cold SKT. Instead, the oceanic LHF anomalies are associated with lower than normal RH over parts of the Atlantic and eastern Pacific in DJF and the western North Pacific and Atlantic as well as the South Pacific in JJA. Enhanced near-surface wind velocities are found over parts of the Southern Hemisphere oceans, in particular along the African east coast in DJF and over the western North Pacific in JJA (Figs. 5g,h). The fact that the anomalies in RH and V10 are mostly weaker than the LHF anomalies points to a combined effect of both low RH and high wind speed causing the enhanced evaporation. It is unlikely that the oceanic anomalies are also affected by a daily cycle, since there is no signal in SKT, and the daily cycle of RH over the ocean is generally very weak (see Dai 2006). Pfahl and Niedermann (2011) studied the origin of oceanic near-surface RH anomalies on daily time scales and found that in the extratropics low RH values are related to equatorward transport and descending air masses, either due to large-scale subsidence or turbulent mixing. Such a descent is likely associated also with high wind velocities near the surface because of downward momentum transport. It can be speculated that, for example, an upper-level trough may induce both strong evaporation due to equatorward advection at its rear and poleward advection in the form of a WCB at its front side (cf. Winschall et al. 2012). Such a scenario differs strongly from the conditions at the more tropical moisture sources, where no specific trigger of enhanced evaporation is required. Future research is warranted to further detail the synoptic-scale conditions linking the WCB ascent and its moisture supply.

The results presented in Figs. 4 and 5 are based on a relatively high uptake humidity threshold of Δqupt = 2 g kg−1 (see section 2b). When this threshold is lowered to 1 or 0.5 g kg−1, the number of grid points where the frequency of WCB moisture uptakes is larger than 1% somewhat increases, but the anomaly patterns are very similar as for Δqupt = 2 g kg−1 (not shown). In the main WCB source regions, the magnitude of the anomalies is slightly reduced for lower threshold values, as the humidity uptakes with a smaller Δq are associated with a weaker anomaly signal.

4. Relevance for precipitation

a. An example WCB

To illustrate our method for attributing precipitation to WCBs, an example of a precipitation event related to a WCB in southern Europe in October 2000 is shown in Fig. 6. This event contributed to a major flooding in southern Switzerland (Bundesamt für Wasser und Geologie 2002). Between 0000 and 0600 UTC 15 October, heavy precipitation occurred in a band over the Mediterranean Sea, northwestern Italy, and southern France, with peak values of more than 30 mm in 6 h (Fig. 6b). The stippling in Fig. 6b indicates where the 6-hourly precipitation exceeded its 99th percentile and is thus classified as extreme (see section 2c). The precipitation occurred in the vicinity of a weak low pressure system over the Gulf of Lions at 0600 UTC 15 October (Fig. 6a) at the front side of a pronounced upper-level trough extending from the British Isles to Tunisia, as shown by the PV contours in Fig. 6b. The curvature of the PV contours gives an indication of cyclonic wave breaking in the upper troposphere. Figure 6a shows the set of WCB trajectories that are related to the precipitation pattern depicted in Fig. 6b. Only those WCB trajectories are shown along which the specific humidity decreased by more than Δqpr = 1 g kg−1 between 0000 and 0600 UTC 15 October (see again section 2c). The blue trajectory segments indicate where this moisture loss took place; trajectory segments before 0000 UTC and after 0600 UTC are shown in dark and light gray, respectively. The precipitation in all grid boxes surrounding the blue trajectory segments is associated with the WCB, as indicated by the dashed purple line in Fig. 6b. Note that this line comprises also grid points with no precipitation at the surface, which is mainly because it defines the envelope of the trajectory segments and the WCB precipitation is assumed to be equally distributed over 6-hourly periods. Several WCB trajectory bundles can be identified in Fig. 6a that ascended directly east of the upper-level trough (and close to the cyclone’s cold front; not shown). These trajectories are connected to the main frontal precipitation band reaching from the African coast to northern Italy. The secondary precipitation maximum close to the cyclone center in southern France relates to a more westerly trajectory bundle that started earlier. The rainfall related to the WCB was mainly of large-scale character, but embedded convection may also have played a role. Some rain over the Mediterranean Sea between the Balearic Islands and Sardinia is not associated with the WCB. This precipitation was mainly due to convection in the cyclone’s cold sector. This example illustrates that WCBs can lead to substantial rainfall at the surface and motivates a climatological investigation of the linkage between WCBs and precipitation.

Fig. 6.

(a) WCB trajectories contributing to the 6-hourly accumulated precipitation in southern Europe between 0000 and 0600 UTC 15 Oct 2000. Trajectory segments with a moisture loss of >1 g kg−1 (6 h)−1 in the respective 6-hourly period are shown in blue. Dark (light) gray lines indicate trajectory segments before 0000 UTC 15 Oct (after 0600 UTC 15 Oct). Note that the displayed 48-h ascent of the trajectories started at different time instants between 0600 UTC 13 Oct and 0000 UTC 15 Oct. The orange contours show sea level pressure at 0600 UTC 15 Oct (contour interval 3 hPa; contours ≥ 1014 hPa dashed). (b) The 6-hourly accumulated precipitation between 0000 and 0600 UTC 15 Oct 2000 (color shading, mm) and potential vorticity on the 330-K isentrope at 0600 UTC 15 Oct [green contours of 1.5 and 2 potential vorticity units (PVU)]. Areas are stippled where the 6-hourly accumulated precipitation exceeds its 99th percentile, which has been calculated separately at each grid point for the whole period 1979–2010. The purple dashed line shows the region affected by WCB precipitation [based on the trajectories in (a)].

Fig. 6.

(a) WCB trajectories contributing to the 6-hourly accumulated precipitation in southern Europe between 0000 and 0600 UTC 15 Oct 2000. Trajectory segments with a moisture loss of >1 g kg−1 (6 h)−1 in the respective 6-hourly period are shown in blue. Dark (light) gray lines indicate trajectory segments before 0000 UTC 15 Oct (after 0600 UTC 15 Oct). Note that the displayed 48-h ascent of the trajectories started at different time instants between 0600 UTC 13 Oct and 0000 UTC 15 Oct. The orange contours show sea level pressure at 0600 UTC 15 Oct (contour interval 3 hPa; contours ≥ 1014 hPa dashed). (b) The 6-hourly accumulated precipitation between 0000 and 0600 UTC 15 Oct 2000 (color shading, mm) and potential vorticity on the 330-K isentrope at 0600 UTC 15 Oct [green contours of 1.5 and 2 potential vorticity units (PVU)]. Areas are stippled where the 6-hourly accumulated precipitation exceeds its 99th percentile, which has been calculated separately at each grid point for the whole period 1979–2010. The purple dashed line shows the region affected by WCB precipitation [based on the trajectories in (a)].

b. Climatology

Figure 7a shows the climatological frequency of WCB precipitation (i.e., the percentage of 6-hourly intervals between 1979 and 2010 during which a grid point was located in the vicinity of a WCB trajectory segment with a moisture decrease of more than Δqpr = 1 g kg−1; see section 2c). The spatial pattern is very similar to the climatological WCB distribution (M14). Frequency maxima of more than 15% are found over South America and southeastern Asia related to an orographically enhanced WCB ascent as well as over the western North Pacific. Over the Atlantic Ocean, the highest values are slightly above 10%.

Fig. 7.

(a) Frequency of WCB precipitation [moisture loss > 1 g kg−1 (6 h)−1] in percent of all 6-hourly intervals in 1979–2010. (b) Percentage of total precipitation amount associated with a WCB. (c) Percentage of the number of extreme precipitation events associated with a WCB.

Fig. 7.

(a) Frequency of WCB precipitation [moisture loss > 1 g kg−1 (6 h)−1] in percent of all 6-hourly intervals in 1979–2010. (b) Percentage of total precipitation amount associated with a WCB. (c) Percentage of the number of extreme precipitation events associated with a WCB.

The percentage of total precipitation associated with WCBs has a similar spatial distribution, but the values are much larger (Fig. 7b). Over parts of southern South America and the western North Pacific near Japan, more than 60% of the precipitation falls in the vicinity of a WCB; 40% of the precipitation can be associated with a WCB over larger parts of the extratropical oceans (e.g., over the North and South Atlantic and the South Pacific). Over land, peak values of more than 30% are reached in the southeastern United States, Australia, eastern China, and parts of Arabia and the Middle East. The latter is particularly noteworthy since WCB precipitation is very rare in this region (cf. Fig. 7a).

With respect to precipitation extremes, WCBs are even more important (Fig. 7c). More than 60% of the precipitation extremes are related to a WCB over large parts of the subtropical and midlatitude oceans, with maxima of more than 90% over the Atlantic, North Pacific, South America, and the Himalayan foothills. Many densely populated regions like Japan, Korea, and eastern China, the southeastern United States, and the Rio de la Plata estuary are strongly affected by extreme precipitation associated with WCBs. In Europe south of the Alps, in California, and in the Middle East, where WCBs are comparatively rare, they are nevertheless associated with up to 50% of the extreme precipitation events. In summary, Fig. 7 shows that WCBs, though relatively rare, contribute substantially to total precipitation and to the majority of the extreme precipitation events in many extratropical regions.

The seasonality of the relevance of WCBs for total and extreme precipitation is largely determined by the seasonal distribution of WCB occurrence (not shown); that is, WCBs are more important for precipitation in winter when they are typically more frequent than in summer. If 3 g kg−1 instead of 1 g kg−1 is chosen for the threshold of the specific humidity decrease Δqpr (see section 2c), only strongly precipitating WCBs are taken into account, and both the frequency of WCB precipitation and the percentage of (extreme) precipitation associated with a WCB are lower (the latter by roughly 10%–20%; not shown). Nevertheless, the spatial patterns are unchanged, and the large difference between the frequency of WCB precipitation and its contribution to total and extreme precipitation still exists. This indicates that the general finding of a very important WCB contribution to precipitation holds independently of the specific choice of the threshold value.

For assessing the combined effect of cyclones and WCBs on extreme precipitation events, Fig. 8 shows the percentage of precipitation extremes associated with either a WCB or a cyclone. The dark (light) gray contours highlight regions where most of the events are related to a WCB (cyclone). Note that the percentage of extremes coinciding with a cyclone in the period 1979–2010 is almost identical as in the shorter period analyzed by Pfahl and Wernli (2012). Cyclones or WCBs cause a substantial fraction of the precipitation extremes in large parts of the extratropics, in particular over the ocean, but also in many continental regions. In the Northern Hemisphere storm tracks, there is some overlap between the areas strongly affected by WCBs and cyclones (dark and light contours in Fig. 8), with the cyclone areas extending farther north and the WCBs affecting more southerly regions, especially in the eastern North Pacific. This meridional displacement is due to the typical origin of WCBs in the warm sector, equatorward of the cyclone center (see section 1). The importance of the combined effect of cyclones and WCBs is evident, for example, over the eastern North Atlantic and the British Isles, where none of the two flow features alone explains more than 70% of the precipitation extremes. In the Southern Hemisphere, the meridional shift between the areas strongly affected by cyclones and WCBs is striking (see again Fig. 8). Cyclones are most important over the Southern Ocean, whereas WCBs are mainly associated with precipitation extremes closer to the equator. The Southern Hemisphere maximum of the percentage of precipitation extremes associated with cyclones or WCBs is found over South America and the South Atlantic, where a region dominated by WCBs is directly connected to the main Southern Ocean storm tracks.

Fig. 8.

Percentage of precipitation extremes associated with either a WCB or a cyclone. The dark gray contour indicates regions where >70% of the extreme events are related to a WCB (cf. Fig. 7); the light gray contour shows where >70% of the extreme events are associated with a cyclone.

Fig. 8.

Percentage of precipitation extremes associated with either a WCB or a cyclone. The dark gray contour indicates regions where >70% of the extreme events are related to a WCB (cf. Fig. 7); the light gray contour shows where >70% of the extreme events are associated with a cyclone.

5. Conclusions

Two aspects of the coupling between WCBs and the atmospheric moisture cycle have been analyzed in this study: the evaporative moisture sources of WCBs and their contribution to total and extreme precipitation. The WCB moisture supply has been investigated with a Lagrangian diagnostic based on backward trajectories. This diagnostic revealed important seasonal differences between the spatial and temporal scales of humidity transport to the WCB starting locations. During winter, the moisture sources are located close to the starting points of the WCB ascent and mainly over the ocean. Most of the humidity evaporates during 5 days prior to the start of the ascent. This close temporal and spatial correspondence indicates that evaporation and WCB ascent often occur in the same dynamical environment and may be triggered by the same flow disturbance (e.g., an upper-level trough). In contrast, continental moisture sources and long-range transport from remote evaporation are more important during the summer season. The atmospheric residence time of the humidity before entering the WCB ascent is often longer than 10 days. Furthermore, the moisture transport during summer has a more pronounced meridional component, associated with important contributions of tropical humidity. Such contributions may be linked to specific tropical moisture export events as identified by Knippertz and Wernli (2010), which occur particularly frequently in the western North Pacific during JJA.

Further analyses have shown that WCB moisture uptakes are often associated with temporal LHF anomalies, especially in extratropical regions with relatively weak seasonal-mean LHF. In these regions, stronger than usual evaporation is required to supply sufficient humidity for a WCB. Over land, the daily cycle of evapotranspiration is supposed to contribute to the LHF anomalies, indicating that continental WCB moisture uptake occurs predominantly during daytime. The oceanic LHF anomalies are caused by a combination of reduced near-surface RH and enhanced wind velocity. RH is more important, for example, over the North Atlantic, and wind speed is more important over the western North Pacific. This shows that specific atmospheric circulation patterns are associated with the strong evaporation, which then feeds back on the circulation by providing moisture for a WCB ascent. This complex interaction should be investigated in more detail in future research and models should be evaluated with respect to this mechanism. As already outlined in the introduction, model errors in the representation of WCB moisture supply may have important consequences for the prediction of the downstream atmospheric flow.

In Part II of this paper, it has been shown that WCBs contribute substantially to total precipitation in many regions in the extratropics and are even more important for precipitation extremes. For instance, over parts of South America, eastern China, Japan, and the southeastern United States, more than 70% of the extreme precipitation events are associated with a WCB. The spatial pattern of WCB-related precipitation extremes complements the pattern of extreme events associated with cyclones (see Pfahl and Wernli 2012), with the regions affected by WCBs typically shifted equatorward compared to the maxima of cyclone-induced precipitation extremes. Other recent studies estimated the contribution of cyclones (Hawcroft et al. 2012) and fronts (Catto et al. 2012) to total precipitation, but it is difficult to quantitatively compare their results to our estimates of total precipitation associated with WCBs due to methodological differences, for instance in the identification of cyclones, and since WCBs and fronts are related but not identical features. Moreover, Hawcroft et al. (2012) and Catto et al. (2012) used daily precipitation observations, which require huge search areas for determining the influence of cyclones or fronts. Here, a direct association of WCBs and precipitation has been provided that does not require assumptions on the size of the search region (with the drawback that we relied on predicted precipitation fields). Nevertheless, from all these studies the important role of synoptic-scale flow features for precipitation in the extratropics has become evident. WCBs, with a frequency of occurrence of mostly less than 10%, are not very abundant features, but still explain a huge percentage of the (extreme) precipitation in many regions. This can be of great relevance for weather forecast: an accurate prediction of WCBs is essential also for the forecast of precipitation extremes. Accordingly, climate models have to properly represent WCBs in order to be able to simulate heavy precipitation events in a correct dynamical context.

The detailed analysis of WCBs with respect to their moisture supply and relevance for precipitation presented in this study contributes to a better mechanistic understanding of these important flow features and of the atmospheric water cycle in general.

Acknowledgments

We thank MeteoSwiss and the ECMWF for providing access to ERA-Interim reanalysis data. E. Madonna and M. Boettcher acknowledge funding by the Swiss National Science Foundation (Project 200021-130079). We are grateful to three anonymous reviewers for their helpful comments. The open-source software package R (R Development Core Team 2013) has been used for producing the analyses and graphics for this study.

REFERENCES

REFERENCES
Arnold
,
F.
,
J.
Schneider
,
K.
Gollinger
,
H.
Schlager
,
P.
Schulte
,
D. E.
Hagen
,
P. D.
Whitefield
, and
P.
van Velthofen
,
1997
:
Observation of upper tropospheric sulfur dioxide- and acetone-pollution: Potential implications for hydroxyl radical and aerosol formation
.
Geophys. Res. Lett.
,
24
,
57
60
.
Boutle
,
I. A.
,
S. E.
Belcher
, and
R. S.
Plant
,
2011
:
Moisture transport in midlatitude cyclones
.
Quart. J. Roy. Meteor. Soc.
,
137
,
360
373
,
doi:10.1002/qj.783
.
Browning
,
K. A.
,
1986
:
Conceptual models of precipitation systems
.
Wea. Forecasting
,
1
,
23
41
.
Browning
,
K. A.
,
1990
: Organisation of clouds and precipitation in extratropical cyclones. Extratropical Cyclones: The Erik Palmen Memorial Volume, C. W. Newton and E. O. Holopainen, Eds., Amer. Meteor. Soc., 129–154.
Bundesamt für Wasser und Geologie
,
2002
: Hochwasser 2000. Berichte des BWG, Serie Wasser 2, 248 pp.
Carlson
,
T. N.
,
1980
:
Airflow through midlatitude cyclones and the comma cloud pattern
.
Mon. Wea. Rev.
,
108
,
1498
1509
.
Catto
,
J. L.
,
C.
Jakob
,
G.
Berry
, and
N.
Nicholls
,
2012
:
Relating global precipitation to atmospheric fronts
.
Geophys. Res. Lett.
,
39
, L10805,
doi:10.1029/2012GL051736
.
Dai
,
A.
,
2006
:
Recent climatology, variability, and trends in global surface humidity
.
J. Climate
,
19
,
3589
3606
.
Dee
,
D. P.
, and
Coauthors
,
2011
:
The ERA-Interim Reanalysis: Configuration and performance of the data assimilation system
.
Quart. J. Roy. Meteor. Soc.
,
137
,
553
597
,
doi:10.1002/qj.828
.
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.
, and
Coauthors
,
2012
:
Oceanic and terrestrial sources of continental precipitation
.
Rev. Geophys.
,
50
, RG4003,
doi:10.1029/2012RG000389
.
Grams
,
C. M.
, and
Coauthors
,
2011
:
The key role of diabatic processes in modifying the upper-tropospheric wave guide: A North Atlantic case-study
.
Quart. J. Roy. Meteor. Soc.
,
137
,
2174
2193
,
doi:10.1002/qj.891
.
Harrold
,
T. W.
,
1973
:
Mechanisms influencing the distribution of precipitation within baroclinic disturbances
.
Quart. J. Roy. Meteor. Soc.
,
99
,
232
251
.
Hawcroft
,
M. K.
,
L. C.
Shaffrey
,
K. I.
Hodges
, and
H. F.
Dacre
,
2012
:
How much Northern Hemisphere precipitation is associated with extratropical cyclones?
Geophys. Res. Lett.
,
39
, L24809,
doi:10.1029/2012GL053866
.
Joos
,
H.
, and
H.
Wernli
,
2012
:
Influence of microphysical processes on the potential vorticity development in a warm conveyor belt: A case study with the limited-area model COSMO
.
Quart. J. Roy. Meteor. Soc.
,
138
,
407
418
,
doi:10.1002/qj.934
.
Knippertz
,
P.
, and
H.
Wernli
,
2010
:
A Lagrangian climatology of tropical moisture exports to the Northern Hemisphere extratropics
.
J. Climate
,
23
,
987
1003
.
Lavers
,
R. A.
,
R. P.
Allan
,
E. F.
Wood
,
G.
Villarini
,
D. J.
Brayshaw
, and
A. J.
Wade
,
2011
:
Winter floods in Britain are connected to atmospheric rivers
.
Geophys. Res. Lett.
,
38
, L23802,
doi:10.1029/2011GL049783
.
Liu
,
W. T.
,
K. B.
Katsaros
, and
J. A.
Businger
,
1979
:
Bulk parameterization of air–sea exchanges of heat and water vapor including the molecular constraints at the interface
.
J. Atmos. Sci.
,
36
,
1722
1735
.
Madonna
,
E.
,
H.
Wernli
,
H.
Joos
, and
O.
Martius
,
2014
: Warm conveyor belts in the ERA-Interim dataset (1979–2010). Part I: Climatology and potential vorticity evolution. J. Climate,
27, 3–26
.
Massacand
,
A. C.
,
H.
Wernli
, and
H. C.
Davies
,
2001
:
Influence of upstream diabatic heating upon an Alpine event of heavy precipitation
.
Mon. Wea. Rev.
,
129
,
2822
2828
.
Pfahl
,
S.
, and
H.
Wernli
,
2008
:
Air parcel trajectory analysis of stable isotopes in water vapor in the eastern Mediterranean
.
J. Geophys. Res.
,
113
, D20104,
doi:10.1029/2008JD009839
.
Pfahl
,
S.
, and
N.
Niedermann
,
2011
:
Daily co-variations in near-surface relative humidity and temperature over the ocean
.
J. Geophys. Res.
,
116
, D19104,
doi:10.1029/2011JD015792
.
Pfahl
,
S.
, and
H.
Wernli
,
2012
:
Quantifying the relevance of cyclones for precipitation extremes
.
J. Climate
,
25
,
6770
6780
.
Pomroy
,
H. R.
, and
A. J.
Thorpe
,
2000
:
The evolution and dynamical role of reduced upper-tropospheric potential vorticity in intensive observing period one of FASTEX
.
Mon. Wea. Rev.
,
128
,
1817
1834
.
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
.
R Development Core Team
, cited
2013
: R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. [Available online at http://www.R-project.org/.]
Schäfler
,
A.
,
A.
Dörnbrack
,
H.
Wernli
,
C.
Kiemle
, and
S.
Pfahl
,
2011
:
Airborne lidar observations in the inflow region of a warm conveyor belt
.
Quart. J. Roy. Meteor. Soc.
,
137
,
1257
1272
,
doi:10.1002/qj.827
.
Schemm
,
S.
,
H.
Wernli
, and
L.
Papritz
,
2013
:
Warm conveyor belts in idealized moist baroclinic wave simulations
.
J. Atmos. Sci.
,
70
,
627
652
.
Sodemann
,
H.
, and
A.
Stohl
,
2013
:
Moisture origin and meridional transport in atmospheric rivers and their association with multiple cyclones
.
Mon. Wea. Rev.
,
141
,
2850
2868
.
Sodemann
,
H.
,
C.
Schwierz
, and
H.
Wernli
,
2008
:
Interannual variability of Greenland winter precipitation sources: Lagrangian moisture diagnostic and North Atlantic Oscillation influence
.
J. Geophys. Res.
,
113
, D03107,
doi:10.1029/2007JD008503
.
Stoelinga
,
M. T.
,
1996
:
A potential vorticity-based study of the role of diabatic heating and friction in a numerically simulated baroclinic cyclone
.
Mon. Wea. Rev.
,
124
,
849
874
.
Stohl
,
A.
,
2001
:
A 1-year Lagrangian “climatology” of airstreams in the Northern Hemisphere troposphere and lowermost stratosphere
.
J. Geophys. Res.
,
106
(D7),
7263
7279
.
Wernli
,
H.
,
1997
:
A Lagrangian-based analysis of extratropical cyclones. II: A detailed case-study
.
Quart. J. Roy. Meteor. Soc.
,
123
,
1677
1706
.
Wernli
,
H.
, and
H. C.
Davies
,
1997
:
A Lagrangian-based analysis of extratropical cyclones. I: The method and some applications
.
Quart. J. Roy. Meteor. Soc.
,
123
,
467
489
.
Wernli
,
H.
, and
C.
Schwierz
,
2006
:
Surface cyclones in the ERA-40 dataset (1958–2001). Part I: Novel identification method and global climatology
.
J. Atmos. Sci.
,
63
,
2486
2507
.
Winschall
,
A.
,
S.
Pfahl
,
H.
Sodemann
, and
H.
Wernli
,
2012
:
Impact of North Atlantic evaporation hot spots on southern Alpine heavy precipitation events
.
Quart. J. Roy. Meteor. Soc.
,
138
,
1245
1258
,
doi:10.1002/qj.987
.