1. Introduction
In the midlatitude ocean regions, extratropical cyclones play an important role for the poleward transfer of heat and moisture, the production of precipitation and clouds, and the planet’s radiative balance (e.g., Tselioudis et al. 2000). The moist processes, such as latent heating within the cyclone warm sector, are themselves dependent on the energy exchange between the ocean and the atmosphere (e.g., Booth et al. 2012). As such, surface latent and sensible heat fluxes are important for the genesis of extratropical cyclones (e.g., Mak 1998), their development (e.g., Kuo et al. 1991), and their intensification (e.g., Hirata et al. 2019; Tsopouridis et al. 2021). Most of the previous work that has aimed to understand the role of the surface fluxes on cyclone intensification has used numerical models to separate the effect the cyclone itself has on the fluxes from the effect the fluxes have on the cyclone (e.g., Kuo et al. 1991; Booth et al. 2012; Haualand and Spengler 2020). These experiments, however, have some limitations: they can only be performed for a small number of cases or for a specific region, and have to rely on how well the model represents various aspects of the atmosphere and, more specifically, moist processes. Therefore, an observational benchmark to evaluate simulations that is focused on the interplay between surface heat fluxes, cyclone strength, and precipitation (e.g., Demirdjian et al. 2022) would be useful.
Satellite observations have proven to be a reliable tool for characterizing cloud or precipitation in cyclones. Satellite-based surface heat flux products are also being developed, which, together, allow a large-scale, large-sample-size, long-term examination of the relationships between moist variables and surface fluxes in cyclones. Unlike with models, it is not possible to determine how the surface heat fluxes are directly interacting with the precipitation and cloud processes within cyclones. However, it is possible to analyze relationships from a purely observational standpoint and contrast them with conclusions drawn from controlled simulations. With the relative strengths and weaknesses of models and observations in mind, we use satellite observations of cloud, precipitation, and surface heat fluxes in cyclones to build near-coincident composites as a reference for numerical simulation studies.
This study focuses on extratropical cyclones found at low latitudes (with a center mostly within ±40°N/S) as 1) this is the region of genesis and preconditioning of the cyclones (e.g., Hoskins and Hodges 2002, 2005); 2) this is where warm conveyor belts tend to start their ascent (Madonna et al. 2014); 3) it constrains the latitudinal variations in temperature and dynamics that affect the cyclones; and 4) this is where surface precipitation retrievals are most reliable for cyclones (Naud et al. 2020). Furthermore, it allows us to use the Cyclone Global Navigation Satellite System (CYGNSS; Ruf et al. 2019) surface heat fluxes, which are less impeded by strong precipitation than other products (Crespo et al. 2019).
Extratropical cyclones are the low pressure component of the midlatitude atmosphere’s response to excessive baroclinic instability (e.g., Holton 1992). Near the surface, they can be viewed in terms of air masses: a warm and moist air mass interacts with a colder and drier air mass. Cold and warm surface fronts form the western and poleward boundaries, respectively, of this warm air region, usually referred to as the warm sector. Satellite observations of clouds and precipitation have been extensively used to examine the sensitivity of these fields to various cyclone-specific characteristics, from cyclone strength, environmental moisture availability, age, or location (e.g., Naud et al. 2006; Field and Wood 2007; Hawcroft et al. 2012; Bodas-Salcedo et al. 2012; Booth et al. 2018). These studies have used compositing and conditional sorting techniques to allow a cyclone-relative perspective and the handling of often sparse or irregular distribution of observations.
The main mechanism for precipitation and cloud production in cyclones is the poleward ascent at the warm front of a region of warm and moist air within the warm sector, called the warm conveyor belt (Browning 1986). In the colder air found in the wake of cold fronts, low-level clouds and drizzle dominate and are produced through shallow convection (Houze 1993). In either sector of the cyclones, the air–sea energy exchanges interact with these cloud and precipitation production mechanisms (e.g., Rudeva and Gulev 2011). To explore the potential link between surface heat fluxes and clouds in cyclones, we employ a database of collocated CYGNSS surface heat fluxes and objectively identified extratropical cyclones (Crespo et al. 2021; Naud et al. 2021). We conditionally sort the cyclones based on the surface heat flux strength in either the region in the wake of the cold fronts, where the fluxes are the strongest, or the region between cold and warm fronts, the warm sector, where fluxes are the weakest or negative. Using 3 years of collocated cyclone observations identified within the ±50° latitude band, we contrast weak- and strong-flux cyclones to examine how variations in flux strength or flux sign translate to precipitation and cloud properties in the cyclones.
2. Datasets and methodology
First, we describe the datasets that are used in the analysis. Then we discuss our methods for selecting and sorting cyclones based on the strength of the surface heat fluxes.
a. Datasets
The analysis focuses on the period from August 2018 to September 2021, which is the period of overlap for data availability from CYGNSS and the precipitation data at the time of writing.
1) CYGNSS surface latent and sensible heat fluxes
A full description of the product and algorithm is given in Crespo et al. (2019). CYGNSS is a constellation of eight satellites that collect global positioning system (GPS) signals reflected by the ocean surface in the forward (i.e., specular) direction. The strength of the signal is a function of the surface roughness, thereby providing information on surface wind speed at the location of the specular points. The method allows wind speed retrievals even in conditions of heavy precipitation. With its 35° inclination orbit and 28° off-nadir antenna pointing angles, the mission provides full coverage over ±38° latitude over the course of 24 h (Ruf et al. 2019).
In this study, we use the CYGNSS flux product of Climate Data Record (CDR), version 1.1, which utilizes the CDR 1.1 CYGNSS surface wind speed product and Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2; Gelaro et al. 2017), for the reanalysis. In the tropical oceans, the CYGNSS latent heat flux (LHF) and sensible heat flux (SHF) were found to be in good agreement with buoy estimates, with differences of −3.5 ± 33.6 W m−2 for LHF and 1.53 ± 6.67 W m−2 for SHF (Li et al. 2022). The fluxes are provided for point locations, that is, the specular points, following the observational setup of the CYGNSS mission. The product covers the period from 1 August 2018 to the present day.
2) Extratropical cyclones database
As in Naud et al. (2021), we use a preexisting extratropical cyclones database. The algorithm used to identify and track the cyclones was developed by Bauer and Del Genio (2006). It searches for local minima in sea level pressure and tracks them in time. The database was started using the ERA-Interim sea level pressure (SLP) fields, but for the most recent years of production, this was replaced with the 6-hourly sea level pressure data from the fifth major global reanalysis produced by ECMWF (ERA5; Hersbach et al. 2020). Because we pair the cyclones with CYGNSS flux retrievals, and CYGNSS was launched in 2017, the subset of cyclone tracks we use here is based on ERA5 sea level pressures.
As described in Naud et al. (2010, 2016), cold- and warm-front locations were added to the cyclone identifications a posteriori. Since the highest spatial resolution reanalysis at the time the work was initiated was MERRA-2, the fronts are located using MERRA-2 temperature, geopotential heights and surface winds data. The method makes use of Hewson’s (1998) potential temperature gradient identification routine; for situations where it fails to identify the cold front, the Simmonds et al. (2012) change in wind direction method is used instead (Naud et al. 2016). This second database of cyclones with fronts is publicly accessible for 2006–21.
Given that the CYGNSS dataset is not spatially uniform, a subset of the cyclone and fronts database is used and selected as follows. The cyclone area is defined as a region of radius 1500 km centered on the point of minimum sea level pressure. If at least one CYGNSS specular data point is found in this area within a ±3-h time window, the 6-hourly cyclone occurrence and associated CYGNSS data are included in the CYGNSS extratropical cyclone (ETC) database (see also Naud et al. 2021 for more details). This database currently covers August 2018–December 2021.
Of importance for cloud and precipitation in extratropical cyclones are the strength of the cyclones and the precipitable water available in the area, as demonstrated in Field and Wood (2007). For strength, Field and Wood had chosen surface winds in their work, but in Naud and Kahn (2015), we had proposed using ascent strength instead as it provides additional information on how much lift is available along the warm conveyor belt to form cloud and precipitation. Therefore, MERRA-2 1-hourly precipitable water (PW) and 500-hPa vertical velocity (ω500) fields coincident in time to the cyclone identifications are collected over a circular area of 1500-km radius centered on the location of the SLP minimum.
3) IMERG total precipitation product
For total precipitation, we use precipitation rates reported in the Level 3, final run, Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (IMERG; Huffman et al. 2017), version 6b files (Huffman et al. 2019). The precipitation rates are reported every 30 min in a 0.1° × 0.1° latitude–longitude grid. A full description of the retrieval algorithm is provided in Huffman et al. (2017). Here, we match the gridded total precipitation rate reported in the files to the cyclones by collecting all grid cells found within the 1500-km radius area defined above for the IMERG file closest in time. This makes the cyclone and precipitation fields nearly coincident in time. This precipitation product performed well for extratropical cyclones when compared to other satellite and ground-based products (Naud et al. 2018). At the time of this writing, the dataset covers the period starting in August 2018 and stops in September 2021.
4) MODIS cloud products
For cloud-top height, fraction, and optical thickness data, we collected the Moderate Resolution Imaging Spectoradiometer (MODIS; Salomonson et al. 1989) Aqua daily, version 6.1, products. The MODIS level 3 daily files (Platnick et al. 2015) present the advantages of providing gridded global data and relatively small file sizes, and have been extensively used for examining cloud properties in extratropical cyclones (e.g., Field and Wood 2007; Otkin and Greenwald 2008; Grandey et al. 2013). Therefore, we elected to use this dataset for the work. The daily files provide gridded retrievals of 1° spatial resolution and include the temporal mean of about two overpasses per location. The mean per grid cell, therefore, is performed for observations acquired about 12 h apart, as well as over all of the 1- or 5-km (depending on the product) MODIS pixels that fall into this cell. Naud et al. (2013) demonstrated that using these daily mean cloud fractions for extratropical cyclones composites produced errors of less than 4% when compared to near instantaneous observations. Here, we use the mean of cloud fraction, cloud-top height, and cloud optical thickness. For the optical thickness, only daylight overpasses are included for the mean. For each cyclone in the database, we associate the mean cloud properties for that day.
b. Methodology
Here we describe the 1) classification of cyclones based on the strength of the surface heat fluxes, and 2) the regridding and compositing of cloud and precipitation fields in the cyclone grid. As mentioned earlier, our goal is to classify the whole CYGNSS-ETC database based on how strong the latent and sensible heat fluxes are in the post-cold-frontal area and in the cyclone warm sector. Therefore, we define four distinct classifications based on the following four metrics: 1) mean latent heat flux in the post-cold-frontal region, 2) mean latent heat flux in the warm sector, 3) mean sensible heat flux in the post-cold-frontal region, and 4) mean sensible heat flux in the warm sector.
1) Cyclone classification
The CYGNSS dataset is available within a latitude band of ±38°, but we focus on cyclones that can be outside of this band as long as they have a significant portion of the warm sector and/or the post-cold-frontal zone with CYGNSS data points.
Our first step is to populate a stereographic grid centered on the cyclone SLP minimum, which has 14° polar and 100-km radial resolution with near-coincident CYGNSS flux retrievals, up to 3000 km from the cyclone center. The stereographic grid offers flexibility for placing the data in a cold-front-relative space for the purpose of classification but is not as optimum for compositing fluxes, because the grid cells are not of equal area. We perform a linear regression on the longitude and latitude points along the cold front to find the approximate direction of the cold front (Fig. 1a). Then we rotate and shift the data in the stereo grid (cf. Naud et al. 2016) such that the data are now in a grid symmetrical across the cold front (Fig. 1b). CYGNSS fluxes are averaged in a semicircle of radius 1500 km to the west of the cold front (up to 90° polar) to obtain the post-cold-frontal mean and same to the east of the cold front to obtain the warm-sector mean. Note that we do not use the warm-front location in our construction of the cold-front-based regridding, which means that, in some cases, the mean flux might include grid cells within the warm-frontal region. Division of regions solely based on cold-front position ensures that the warm-sector and post-cold-frontal areas are of similar size.
Example of CYGNSS sensible heat flux (a) projected in a cyclone centered stereographic grid and then (b) after a rotation and lateral shift into a cold-front-centered stereographic grid. In (a) the red line shows the cold-front location for this cyclone, the dashed line is a linear regression in longitude–latitude of the cold-front location, and the black stars delineate the 1500-km-radius area selected for (b). In (b) the vertical dashed line shows the location of the regression line, and the horizontal line is perpendicular at the latitude of the low. The two lower quadrants are used to average the surface heat fluxes: post–cold frontal to the left (PCF) and warm sector to the right (WS). This cyclone occurred at 0600 UTC 11 Jan 2018 at latitude −35.83° and longitude −89.78° and has been flipped such that south is at the top of each panel.
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
In addition to calculating the mean flux, we also save the total number of CYGNSS specular points that fall in each subregion. Cyclones with a very small number of specular points in either of the two cyclone sectors might be characterized by a mean flux that is not representative of the actual flux. Therefore, we need to define a threshold for the minimum allowable number of CYGNSS data points in order to reliably classify a cyclone. Too high of a threshold and the total number of cyclones that can be classified is too small to provide meaningful statistics; too low of a threshold entails too many cyclones that are potentially misclassified.
To reach a decision, we examined the distribution of mean latent and sensible heat fluxes in the post-cold-frontal and warm-sector quadrants when changing the number of data points available to calculate the mean (Fig. 2). The quadrants in question are the semicircles of radius 1500 km and extend 90° on each side of the cold fronts defined earlier (Fig. 1b). For each quadrant separately, we sort the cyclones using the number of data points from low to high. Then we divide the population into five subsets of equal number and examine the histograms obtained for each bin, from the first 20% of the cyclone population with the lowest number of CYGNSS data points to the top 20% with the highest numbers. We use as reference the top 20% and examine how much the less-populated subsets deviate from this reference for the distribution of fluxes. The less-populous subsets tend to have more cases with very low or negative mean fluxes as well as more cases with very large fluxes. The population that includes the top 60%–80% of numbers of CYGNSS points has characteristics that are very similar to the those in the top 20%, so we choose to limit the minimum number of cases to within this subset. If we impose a minimum of 2000 data points (dotted lines in Fig. 2), we obtain distributions of mean fluxes close to the top 20% population while increasing the sample size. With this threshold, we keep 3689 cyclones with at least 2000 data points in the post-cold-frontal region (top 30% of all ETCs) and 5325 cyclones with at least 2000 data points in the warm sector (top 37%). For example, the cyclone in Fig. 1 has enough data points in both sectors with this threshold. Based on previous work (Naud et al. 2021), these sample sizes allow meaningful statistics.
Histograms of mean CYGNSS (a),(c) latent and (b),(d) sensible heat fluxes averaged in the (a),(b) post-cold-frontal region and (c),(d) warm sector of all cyclones observed with CYGNSS between August 2018 and September 2021. Each solid line is obtained using a subset of the CYGNSS-ETC database, according to the number of CYGNSS data points available in each cyclone sector from 1 to the maximum number available (23 786 data points for the post-cold-frontal sector and 30 964 for the warm sector), sorted into five equal size bins (from first 20% to top 20%). The dotted line is the histogram obtained when imposing a minimum of 2000 data points per sector.
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
Using this threshold, we sort the cyclones using mean fluxes in the following four classifications: 1) LHF in the post-cold-frontal region (LHF-PCF), 2) SHF in the post-cold-frontal region (SHF-PCF), 3) LHF in the warm sector (LHF-WS), and 4) SHF in the warm sector (SHF-WS). Then we divide cyclones in each of the four classifications into three subsets of equal numbers of cyclones to obtain weak-, moderate-, and strong-positive-flux categories. As reported in Table 1, mean fluxes in the weak, moderate, or strong category will be overall lower for the warm-sector than for the post-cold-frontal region, consistent with the much stronger fluxes in the latter than former sector. Finally, cyclones in a given classification can be found in the same flux strength category in another, especially for classifications based on the same quadrant but different flux type, or for LHF in opposite quadrants (Table 2).
Range of the mean latent and sensible heat fluxes (F) per sector for each of the three categories for all four classifications.
Number of cyclones per weak, moderate (Mod), and strong category shared between two distinct classifications (Classif). The largest number of cyclones among the three categories is highlighted in boldface.
There are a small number of cases with a negative mean flux in either of the quadrants for LHF-PCF (3), LHF-WS (42), and SHF-PCF (124). These numbers are too small to consider adding a new category, so these cases are not used in this study. However, for SHF-WS, 1273 cyclones have a negative mean SHF in the warm sector. Therefore, for this classification alone, we create a new category that includes cases with a negative flux. Note that a negative flux implies a transfer of energy from the atmosphere into the ocean as opposed to from the ocean to the atmosphere, as for the categories in Table 1. To contrast cloud properties and precipitation with a negative SHF in the warm sector to cases with a positive flux, we create a new category of positive, warm-sector, mean SHF cases tailored to match in absolute value the magnitudes of the flux strength distribution of the negative flux cases. Starting with the negative mean SHF cases, we count the number of ETCs that fall in ±0.5 W m−2 wide bins from –1 to –35 W m−2. For the same bins in absolute flux value, we randomly select positive SHF-WS ETCs such that we have the same number of cases for each SHF bin. This ensures that while the sign of the SHF changes, the magnitude does not overall. For this analysis, we will refer to the two new categories as SHF-WS negative and SHF-WS–positive matched.
2) Data regridding and compositing
To composite fluxes, PW, ω500, precipitation, and cloud properties for each cyclone classification, we define a rectangular grid centered on the cyclone minimum in SLP, composed of equal-area cells defined based on their zonal and meridional distance to the center in kilometers. This type of grid is better suited for precipitation or surface heat flux compositing than is the stereographic grid used earlier, as it preserves the area of each grid cell, thereby avoiding potential distortions around the edges or near the center. The rectangular grid expands to ±1500 km in both directions and each grid cell is 100 km × 100 km. For each cyclone, we collect fluxes, PW, ω500, precipitation, and cloud property data points and populate the grid cells based on their distance to the cyclone center. Then, to produce composites, we accumulate all data points for all cyclones and calculate the mean per grid cell. For precipitation, we include 0 mm h−1 instances in the mean; for cloud fraction, we include pixels with no cloud. However, for both cloud-top height and optical thickness that are relevant only when clouds are present, we only consider pixels for which cloud fraction is strictly greater than zero.
To assess whether differences between composites are significant, we randomly select 1000 cyclones and composite each variable for this subset. This is repeated 200 times, giving 200 composites to calculate a standard deviation for each grid cell. The difference between the different flux strength category is then only considered if it exceeds this standard deviation.
3. Characteristics of the cyclones classified with the surface heat fluxes
In this first results section, we explore the characteristics of the cyclones for each classification, starting with the fluxes themselves, then the cyclone locations, and, finally, their properties.
a. Composites of surface heat fluxes per classification
In our first analysis of the cyclone-relative fluxes, we composite the surface heat fluxes in the cyclones that are used for each of the four classifications of Table 1 (i.e., LHF for LHF-PCF and LHF-WS, SHF for SHF-PCF and SHF-WS). For each classification, we separate the cyclones into the three positive flux categories and find consistent changes among these categories based on the choice in classification (Fig. 3). Given that CYGNSS data are not readily available beyond ±38° latitude, the composites become noisy for the cyclone area poleward of 500 km, despite only considering fluxes in grid cells that contain at least 100 data points. Additionally, since the composites include both hemispheres, the Southern Hemisphere cyclones are flipped along the north–south direction to ensure the polar side is at the top of each panel (as in Fig. 1).
Composites of CYGNSS (a)–(f) latent and (g)–(l) sensible heat fluxes in cyclones classified according to the strength of the flux in post-cold-frontal and warm sectors: (a)–(c) LHF-PCF, (d)–(f) LHF-WS, (g)–(i) SHF-PCF, and (j)–(l) SHF-WS.
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
We can verify that when the cyclones are selected based on the fluxes in the post-cold-frontal region of the cyclones, the three flux strength categories show a significant contrast in the flux in the western half of the cyclone composites (Figs. 3a–c,g–i). In fact, the difference between strong- and weak-flux categories (Figs. 4a,c,e,g) indicates that in the warm sector (eastern half), the fluxes are unchanged or weaker in the strong- versus weak-flux categories. The contrast is more significant in the warm sector when the classification uses warm-sector fluxes (Figs. 3d–f,j–l). The composites also reveal that strong warm-sector fluxes are accompanied by strong post-cold-frontal fluxes (Figs. 4b,d,f,h).
Difference in cyclone centered composites of (a)–(d) LHF and (e)–(h) SHF between strong- and weak-flux categories for (a),(e) LHF-PCF, (b),(f) LHF-WS, (c),(g) SHF-PCF, and (d),(h) SHF-WS classifications.
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
Overall, we find distinct contrasts in the flux differences when we consider which region of the cyclone is used and much less contrast when we use latent or sensible heat flux. Because of the strong impact of the climatological mean in surface heat fluxes and the actual cyclone strength on the fluxes in the cyclones, we next explore where the cyclones are found for each category and what their properties are.
b. Preferred locations of the cyclones classified with the surface heat fluxes
To examine where the cyclones are found for each category, we first check the location of the cyclone centers (and not where the fluxes are sampled) for the two extreme categories, weak- and strong-flux cyclones. Overall, cyclones with enough CYGNSS data points in the post-cold-frontal region are, not surprisingly, positioned such that enough open ocean area is available to the west of the center, explaining the clearing along the east coasts of the continents (Fig. 5c). In contrast, cyclones with enough data points in the warm sector (Fig. 6c) can be found along these eastern coast lines, the largest shift being found in the Atlantic, in both hemispheres. For all classifications, there are more cyclones in the Southern Hemisphere; within each hemisphere, the distribution across categories and classifications is provided in Table 3 and in Figs. 5 and 6.
Location of ETC centers for weak- (blue shading) and strong- (orange lines) flux categories for classifications (a) LHF-PCF and (b) SHF-PCF, as well as (c) for all cyclones with at least 2000 CYGNSS data points in the post-cold-frontal area. The dotted and solid lines are the number of cyclones, contoured every 5, from a minimum of 5.
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
Location of ETC centers for weak- (blue shading) and strong- (red lines) flux categories for classifications (a) LHF-WS and (b) SHF-WS, as well as (c) for all cyclones with at least 2000 CYGNSS data points in the warm sector. The dotted and solid lines are the number of cyclones, contoured every 5, from a minimum of 5.
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
Number of cyclones with weak and strong surface heat fluxes and total for Northern (NH) and Southern (SH) Hemispheres for all four classifications.
For the LHF-PCF classification (Fig. 5a), the weak-flux category is more prevalent in the Southern Hemisphere and the strong-flux category is more prevalent in the Northern Hemisphere. Cyclones with the largest fluxes tend to be located in the regions of western boundary currents, in particular the Gulf Stream region for the Northern Hemisphere or the East Australian Current for the Southern Hemisphere (note the larger number of strong-flux cyclones than weak-flux cyclones there). Cyclones with the weakest fluxes are found in the South Pacific, with a large number of cyclone centers along the Andes, for which the post-cold-frontal area is found over the relatively cold waters of the eastern South Pacific. For the SHF-PCF cyclones (Fig. 5b), both weak- and strong-cyclone categories are relatively evenly represented per hemisphere. The North Atlantic cyclones show a slight shift between strong-flux cases close to the American continent and weak-flux cases closer to Europe.
For the LHF-WS classification (Fig. 6a), the impact of the warm ocean currents on the strong-flux cyclone locations is stronger than seen for the LHF-PCF classification, with most strong-flux cyclones in the Gulf Stream and Kuroshio regions. While there are more cyclones overall in the database in the Southern Hemisphere, the number of stronger-flux cyclones in the Northern Hemisphere is greater. In contrast, the weak-flux cyclones are found predominantly in the North Pacific and South Atlantic Oceans, with relative numbers that are similar to the overall distribution across hemispheres. This suggests that the collocation of cyclones and ocean western boundary currents greatly increases the likelihood of generating strong LHF in the warm sector. For the SHF-WS classification (Fig. 6b), weak-flux cyclones are equally represented in the Northern Hemisphere as their strong-flux counterparts, similar to the SHF-PCF case.
Overall, the maps of cyclone location based on classification strength show that the impacts of western boundary currents on cyclone fluxes are similar to what is observed in climatologies of surface latent and sensible heat fluxes (e.g., Hartmann 2016, Fig. 4.19). They also reveal that, given the focus on low-latitude cyclones, there are limited changes in the latitude of weak- and strong-flux cyclones.
c. Properties of the cyclones classified with the surface heat fluxes
We next examine how the properties of the cyclones differ between flux-strength categories. Contrasts in cold- and warm-front locations across categories are difficult to appreciate given the variability within each category (Fig. S1 in the online supplemental material). Instead, for each flux classification, we construct cyclone-centered composites for all three categories of flux strength of the following cyclone characteristics: mean PW in the cyclone 1500-km-radius area, and mean 500-hPa vertical velocity.
Cyclones with weak surface latent (Fig. 7a) or sensible heat (Fig. 7e) fluxes in the post-cold-frontal region are not just weaker than their strong-flux counterparts (Figs. 7c,g); they also appear a lot more compact, with a footprint within 1000 km of the cyclone center, while the strong-flux category ETCs expand further than the 1500-km-radius region we consider. The main difference between classifying based on PCF LHF or SHF is found for PW: the PW contrast between strong- and weak-LHF ETCs is positive to the east and negative to the west (Fig. 7d), while it is negative everywhere when using SHF instead (Fig. 7h).
Cyclone-centered composites of (columns 1–3) PW (colored contours in 2-mm increments) and 500-hPa vertical velocity (solid line contours in 2 hPa h−1 increments where negative, dashed line in 1 hPa h−1 increments where positive) for weak-, moderate-, and strong-flux categories, respectively, and (column 4) difference between the strong- and weak-flux categories in PW (colored contours above sample size variability) and vertical velocity (solid where negative, dashed contours where positive in 1 hPa h−1 increments) for ETCs classified with (a)–(d) LHF-PCF, (e)–(h) SHF-PCF, (i)–(l) LHF-WS, and (m)–(p) SHF-WS.
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
In contrast, cyclones classified based on surface flux strength in the warm sector show very little differences in terms of ascent or subsidence strength between flux-strength categories (Figs. 7i–p). Again, when considering LHF-WS, PW contrasts resemble those found when changing LHF strength in the PCF (Fig. 7l vs Fig. 7d), while changing SHF-WS strength is accompanied by a lower PW in strong-flux than in weak-flux cases (Fig. 7p).
Field and Wood (2007) demonstrated that precipitation rates and high cloud cover in the warm conveyor belt region both depend on the strength of the cyclone and the mean environmental PW. Therefore, we next examine how competing effects of PW and ascent strength, as well as differing flux strengths, impact clouds and precipitation in flux-based classified cyclones.
4. Contrast in cloud and precipitation in cyclones based on the surface heat fluxes strength from the ocean
To help provide a detailed picture of how precipitation, cloud fraction, cloud-top height, and cloud optical thickness differ between cyclones selected based on the surface heat flux strength, we use both individual composites for each of the three flux-strength categories (Figs. 8–11) and difference composites between strong- and weak-flux subsets (Fig. 12).
Cyclone-centered composites of (a)–(c) IMERG precipitation, and MODIS (d)–(f) cloud-top height, (g)–(i) cloud fraction and (j)–(l) cloud optical thickness for cyclones with (left) weak, (center) moderate, and (right) strong LHF in the post-cold-frontal region (LHF-PCF classification).
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
Cyclone-centered composites of (a)–(c) IMERG precipitation and (d)–(f) MODIS cloud-top height, (g)–(i) cloud fraction, and (j)–(l) cloud optical thickness for cyclones with (left) weak, (center) moderate, and (right) strong SHF in the post-cold-frontal region (SHF-PCF classification).
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
Cyclone-centered composites of (a)–(c) IMERG precipitation, and (d)–(f) MODIS cloud-top height, (g)–(i) cloud fraction, and (j)–(l) cloud optical thickness for cyclones with (left) weak, (center) moderate, and (right) strong LHF in the warm sector (LHF-WS classification).
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
Cyclone-centered composites of (a)–(c) IMERG precipitation, and (d)–(f) MODIS cloud-top height, (g)–(i) cloud fraction, and (j)–(l) cloud optical thickness for cyclones with (left) weak, (center) moderate, and (right) strong SHF in the warm sector (SHF-WS classification).
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
Difference in cyclone-centered composites of (a)–(d) precipitation rates, (e)–(h) cloud-top height, (i)–(l) cloud fraction, and (m)–(p) cloud optical thickness between strong- and weak-flux categories for (a),(e),(i),(m) LHF-PCF; (b),(f),(j),(n) SHF-PCF; (c),(g),(k),(o) LHF-WS; and (d),(h),(l),(p) SHF-WS classifications. Solid contours represent the composites for the weak-flux category.
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
a. Contrast in composites of cloud-top height and precipitation for different flux strength classifications
To help describe the contrasts in precipitation and cloud-top height, we distinguish two sectors of the cyclones: 1) the comma region that coincides with the area of ascent and relatively large PW, and 2) the western half of the cyclones where subsidence and relatively low PW dominate (see Fig. 7). As illustrated in Figs. 8–11, the ascent region displays the heaviest precipitation and highest cloud top in the cyclones, while the subsidence dominated area is populated by lightly precipitating shallow clouds.
For the PCF flux classifications (Figs. 8 and 9), strong-flux cyclones have heavier precipitation and higher clouds in the comma region, with lighter precipitation and shallower clouds in the subsidence zone than their weak-flux counterparts (see also Figs. 12a,b,e,f). This is easily explained by the contrast in cyclone strength and PW (Fig. 7): the stronger the ascent and larger PW in the comma region, the heavier the precipitation and the higher the clouds. A stronger subsidence will be accompanied by a shallower planetary boundary layer and stronger inversion, and, therefore, relatively shallower clouds, while drier conditions will lead to lighter precipitation.
For the warm-sector flux classifications (Figs. 10 and 11), the contrast in precipitation and cloud-top height is of similar sign and magnitude in the subsidence-dominated sector as what is seen for PCF-based flux classifications, presumably because even though the subsidence strength does not change much across categories, strong-flux cyclones display lower PW than their weak counterparts. This is not the case in the comma region for which the choice of flux (LHF vs SHF) matters. In the comma region, if cyclones have strong LHF in the warm sector, precipitation is slighter heavier and cloud top is higher than in weak-LHF cyclones. However, strong SHF in the warm sector is accompanied by lighter precipitation in the comma head as well as shallower clouds, compared to weak-SHF cyclones (Figs.12c,d,g,h). For these warm-sector flux classifications, the strength of the cyclones does not change much between weak- and strong-flux cyclones; therefore, the contrast in PW plays a larger role. So precipitation and cloud-top height contrasts are consistent with PW contrasts in the comma head between strong and weak cyclones: they are positive for the LHF-based classification, versus negative for the SHF-based classification.
b. Contrast in composites of cloud fraction and optical thickness for different flux-strength classifications
When examining the distributions of cloud fraction and optical thickness in Figs. 8–11, a clear contrast in the region of the comma head can be seen, so we focus on this area first (i.e., where the ascent strength is maximum). For the classifications based on the PCF fluxes, the cloud fraction is larger in cyclones with strong fluxes than in those with weak fluxes (Figs. 8g–i and 9g–i); this is expected, given that cloud fraction depends on ascent strength as much as precipitation or cloud-top height. For warm-sector-based classifications, strong-SHF cyclones have lower cloud fractions in the comma head than their weak-SHF counterparts (Fig. 12l), consistent with the contrast in PW (Fig. 7). The contrast in cloud fraction for strong- versus weak-LHF cyclones (Fig. 12k) is similarly distributed as what is seen for cloud-top height (Fig. 12g), but the band of higher cloud tops along the western edge of the comma head displays a negligible contrast in cloud fraction. Nevertheless, the contrasts in cloud fraction when considering warm-sector-based classification appear consistent with the contrasts in PW (Fig. 7).
However, outside of the comma-head region, strong LHFs or SHFs in the PCF tend to be accompanied by larger cloud fractions in most of the cyclone area, and strong fluxes in the warm sector are accompanied by lower cloud fractions. This is consistent with the contrasts in vertical velocity for PCF-based classifications and in PW for the others for the ascending region of the cyclones. However, in the area where subsidence dominates, cyclones with strong warm sector and PCF LHF have lower cloud fraction than cyclones with weak LHF (Figs. 12i,k), while the opposite is true for cyclones with strong versus weak SHF (Figs. 12j,l). In this subsidence dominated area, the contrast in subsidence strength is of the same order regardless of which flux we use and, for both SHF- and LHF-based classifications, strong-flux cyclones display lower PW than weak-flux cyclones. Therefore, in this case, the contrast in cyclone properties does not explain the contrast in cloud fraction. Instead, we turn to the definition of LHF and SHF from the bulk formula. With the cyclone strength being relatively unchanged between strong- and weak-flux cases, we can assume wind speeds also are unchanged. For LHF, the dominating factor will be the contrast in air specific humidity, that is, strong-flux cases will have relatively lower air specific humidity values than the weak-flux cases, implying that the strong-flux conditions will make cloud formation and maintenance harder than in the wetter weak-flux cases. Therefore, lower cloud fraction in the subsidence area for strong-LHF cases is to be expected. For SHF, though, the main contrast between strong- and weak-flux cases will be lower air temperature in the former. While the air might also be drier, it remains that lower temperatures will help condensation and, therefore, cloud formation and maintenance. Therefore, larger cloud fraction in the subsidence area for strong-SHF cases can be explained.
Turning now to optical thickness, we find relatively similar contrasts across all four classifications (Figs. 8–11j–l and 12m–p): In the comma head, the contrast is virtually null, possibly because clouds are so opaque that it might be difficult for a radiometer such as MODIS to sense small changes in reflected solar radiation. Everywhere else, strong-flux cases have more opaque clouds than weak-flux cases, with the magnitude of the contrast being larger for the PCF-based classifications than warm-sector-based ones. This might be because enhanced fluxes lead to more evaporation and, therefore, more water available in the clouds when they form.
To summarize, cloud-top height and precipitation in cyclones selected for the strength of surface heat fluxes tend to be strongly related to the strength of the ascent and the environmental PW in the cyclone. Cloud cover follows similar relationships in the comma head, but in the subsidence area of the cyclone, whether latent or sensible heat flux is considered matters. Finally, cloud optical thickness is always larger for strong- than weak-flux cyclones, suggesting it is more closely related to air–sea interactions than the other cloud properties. Together, these contrasts in the subsidence region where shallow convection dominates suggest that strong surface heat fluxes might enhance condensation, but morphology might change depending on whether latent or sensible heat flux is strong.
Up to now, we have explored the contrasts in clouds and precipitation in cyclones based on the strength of the surface heat fluxes coming out of the ocean. Next, we explore these relationships when we keep the flux strength similar but contrast the direction of the sensible heat flux in the warm sector instead.
5. Contrast in cloud and precipitation in cyclones based on the direction of the surface sensible heat fluxes
As explained in section 2, there are enough cyclones with a negative sensible heat flux in the warm sector to build composites. By contrasting composites obtained for all negative flux cases to matched positive flux cases, we can examine how clouds and precipitation in the cyclone differ when SHF is negative or positive. The location of these negative SHF-WS cases (not shown) is more or less similar to what is found for positive SHFWS cases (cf. Fig. 6).
In our cases with negative SHF, warm moist air in the warm sector of the cyclones has been advected over cooler water and, in this case, the atmosphere fluxes energy into the ocean. This is clearly illustrated in Fig. 13, in which the cyclone-centered composites of PW and 500-hPa vertical velocity for the two opposite categories of SHF-WS negative and SHF-WS–positive matched. It indicates that the region of ascent is more vigorous for negative SHF cases, with larger PW in the tail of the ascent region, while PW is less in the poleward and westward areas of the cyclones, with little contrast in subsidence strength.
Cyclone-centered composites of PW (colored contours every 2 mm) and 500-hPa vertical velocity (solid contours every 2 hPa h−1 where negative and dashed every 1 hPa h−1 where positive) for (a) ETCs with SHF < 0 W m−2 in the warm sector (SHF-WS negative) and (b) SHF > 0 W m−2 (SHF-WS–positive matched); and (c) difference in PW (colored contours every 1 mm) and 500-hPa vertical velocity (solid where negative, dashed where positive, every 1 hPa h−1).
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
CYGNSS LHF and SHF composite differences (Figs. 14a,b) show lower fluxes in the ascent region for negative SHF cases, as expected; we find a slightly positive SHF difference in the western half. Consistent with the contrasts in PW and ascent strength (Fig. 13c), when SHF is negative, precipitation is greater in the comma region, both head and tail (Fig. 14c), as well as in cloud-top heights, (Fig. 14c), cloud fraction, and optical thickness (Figs. 14e,f).
Difference in cyclone-centered composites of (a) CYGNSS LHF, (b) CYGNSS SHF, (c) IMERG precipitation rate, (d) MODIS cloud-top height, (e) cloud fraction, and (f) cloud optical thickness between negative and positive SHF-WS flux categories. Solid contours represent the composites for SHF-WS–positive matched.
Citation: Journal of Climate 36, 16; 10.1175/JCLI-D-22-0600.1
These contrasts are comparable to contrasting weak- and strong-SHF-WS positive fluxes for precipitation and cloud-top heights (Figs. 12d,h), but not equivalent for cloud fraction or optical thickness (Fig. 12l vs Fig. 14e or Fig. 12p vs Fig. 14f). Interestingly, we do not see a clear distinction between the contrasts in the comma region and those in the western half of the cyclones as we saw earlier for differing strength of positive SHF. When sensible heat fluxes in the warm sector are negative (i.e., there is a transfer of heat from the atmosphere into the ocean), the cyclones tend to have more extensive and deeper clouds and greater precipitation in the entire cyclone footprint, compared to cases with positive fluxes, or with a net transfer of heat in the atmosphere from the ocean.
6. Conclusions and discussion
Using observed MODIS cloud properties and IMERG precipitation rates in extratropical cyclones, we contrast clouds and precipitation in cyclones with 1) strong versus weak CYGNSS surface heat fluxes of similar sign or 2) positive versus negative sensible heat fluxes in the warm sector of similar magnitude. We chose to separately examine the contrasts in cloud and precipitation in the western half of the cyclones, where (typically) low-level clouds and light precipitation dominate, from the comma region where high clouds are abundant and precipitation rates are large.
In the comma region, contrasts in precipitation and cloud properties between strong- and weak-flux cases are consistent with contrasts in cyclone strength and PW. Strong fluxes in the post-cold-frontal region tend to occur in strong cyclones and, in turn, strong cyclones display vigorous poleward moisture transport and lift that favor heavy precipitation and extensive cloud fields. Therefore, strong-PCF flux cyclones display heavier precipitation, higher cloud tops, and more extensive cloud fraction than weak-flux cases, regardless of which flux is used for the classification (i.e., LHF or SHF). If, instead, the mean flux in the warm sector is used, the contrast in cyclone PW dominates and it depends on which flux is used: strong-LHF cyclones have heavier precipitation and more clouds in the comma region than weak-LHF cyclones, but strong-SHF cyclones have less. In the comma head, optical thickness changes little between categories for all four classifications, presumably because it is so large that variations are not easily measurable with MODIS.
In the western half of the cyclones, where subsidence dominates, contrasting strong- versus weak-flux categories reveals a systematic decrease in precipitation and cloud-top height across all four classifications. However, cloud fraction contrasts depend on which flux is used: strong-LHF cyclones display lower cloud fraction than weak-LHF cyclones, while strong-SHF cyclones display larger cloud fraction than their weak counterparts. This is presumably inherent to the contrast in humidity or temperature that define latent and sensible heat fluxes. Lower air humidity in strong-, rather than weak-, LHF conditions entails less moisture available to form or maintain clouds and, therefore, lower cloud fraction, while lower air temperature in strong- than weak-SHF conditions benefits condensation and, therefore, larger cloud fractions. Optical thickness contrasts are larger for PCF-based flux classification, but the sign is the same for all classifications: clouds are more opaque for larger flux cases, possibly because enhanced evaporation in strong-flux conditions helps clouds retain larger liquid or ice water contents.
Contrasting cyclones with negative and positive SHF in the warm sector confirms that negative SHF in the warm sector occurs when a poleward flow of moist and warm air is accompanied by deep, thick, and extensive precipitating clouds. The SHF-WS–negative cyclones are coincidentally those with the largest cloud cover in the warm sector. As such, the sensible heat flux tends to respond to the atmospheric conditions, and little can be said about how they could impact the cyclone or the cloud properties
Previous studies using idealized numerical simulations of extratropical cyclones have shown that strong latent or sensible heat fluxes in the cold sector (PCF) destabilize the boundary layer, enhancing the occurrence of shallow convection and precipitation production (e.g., Gutowski and Jiang 1998; Boutle et al. 2010). Our results imply that this shallow convection occurs, given the relatively low cloud-top heights and large optical thickness. However, our results also imply that for the strongest-LHF cases, either 1) the convection is suppressed, or 2) some other factor, such as low relative humidity, limits the number of clouds observed in these cases.
Booth et al. (2012) showed that both sensible and latent heat fluxes in the warm sector play a role in the cyclone’s development phase, but the latent heat flux seems to have a larger role for further intensification. Furthermore, Demirdjian et al. (2022) found a link between strong latent heat fluxes in the warm-frontal region and the comma-head precipitation maximum. These studies are consistent with the relatively larger comma-head precipitation and cloud-top heights for strong than weak LHF-WS despite slightly suppressed cloud fractions.
The pathway for the supply of moisture into the warm conveyor belt has been expanded in recent work (Dacre et al. 2019; Haualand and Spengler 2020; Bui and Spengler 2021; Papritz et al. 2021). In particular, these studies show that only in mature or decaying cyclones can strong LHF and SHF near the cyclone center help provide moisture in a cyclonic flow around the cyclone center.
While we are not able to attribute any physical relations among the surface heat fluxes, cyclone properties, cloud, or precipitation, the observations composited here tend to agree with previous work based on idealized case simulations. In any case, they provide a useful constraint for model evaluation. Since we elected to focus primarily on low-latitude cyclones, while we do have cyclones past their peak intensity, more than half eventually intensify and propagate out of the area we focus on. The next step will be to track all of these cyclones and examine how they eventually evolve in strength but also in their overall lifetime precipitation production, given the surface heat flux strength as documented here. These observations can also help when examining the role of strong surface heat fluxes for a downstream cyclone’s strength and precipitation, in conjunction with numerical simulations to help establish causal links.
Acknowledgments.
The work is funded by the NASA programs NNH20ZDA001N-CYGNSS and NNH21ZDA001N-PMMST. J.F.B. and C.M.N. were partly funded through CYGNSS Grant 80NSSC21K1006, and C.M.N. received additional funding through Grants CYGNSS 80NSSC21K1470 and PMM 80NSSC22K0602. D.J.P.’s contributions to this study were carried out on behalf of the Jet Propulsion Laboratory, California Institute of Technology, under a contract with the National Aeronautics and Space Administration. The authors thank three anonymous reviewers for their constructive comments, which improved the manuscript.
Data availability statement.
MODIS L3 daily products (Platnick et al. 2015) are described at https://atmosphere-imager.gsfc.nasa.gov/products/daily and are available at https://ladsweb.modaps.eosdis.nasa.gov/search/order/1/MYD08_D3–61. IMERG final precipitation products (Huffman et al. 2019) are available at the GES DISC (https://disc.gsfc.nasa.gov/datasets/GPM_3IMERGHH_06/summary?keywords=%22IMERG%20final%22). The cyclone database is documented and available at https://data.giss.nasa.gov/storms/obs-etc web site, and is hosted at the NCCS data portal (https://portal.nccs.nasa.gov/datashare/Obs-ETC/Fronts-ETC). The CYGNSS level 2 surface heat flux product of version CDR 1.1 is available through NASA’s PO.DAAC at https://podaac.jpl.nasa.gov/CYGNSS?sections=data. The MERRA-2 products are available through NASA GES DISC (https://disc.gsfc.nasa.gov/datasets?project=MERRA-2).
REFERENCES
Bauer, M., and A. D. Del Genio, 2006: Composite analysis of winter cyclones in a GCM: Influence on climatological humidity. J. Climate, 19, 1652–1672, https://doi.org/10.1175/JCLI3690.1.
Bodas-Salcedo, A., K. D. Williams, P. R. Field, and A. P. Lock, 2012: The surface downwelling solar radiation surplus over the Southern Ocean in the Met Office model: The role of midlatitude cyclone clouds. J. Climate, 25, 7467–7486, https://doi.org/10.1175/JCLI-D-11-00702.1.
Booth, J. F., L. Thompson, J. Patoux, and K. A. Kelly, 2012: Sensitivity of midlatitude storm intensification to perturbations in the sea surface temperature near the Gulf Stream. Mon. Wea. Rev., 140, 1241–1256, https://doi.org/10.1175/MWR-D-11-00195.1.
Booth, J. F., C. M. Naud, and J. Jeyaratnam, 2018: Extratropical cyclone precipitation life cycles: A satellite-based analysis. Geophys. Res. Lett., 45, 8647–8654, https://doi.org/10.1029/2018GL078977.
Boutle, I. A., R. J. Beare, S. E. Belcher, A. R. Brown, and R. S. Plant, 2010: The moist boundary layer under a mid-latitude weather system. Bound.-Layer Meteor., 134, 367–386, https://doi.org/10.1007/s10546-009-9452-9.
Browning, K. A., 1986: Conceptual models of precipitation systems. Wea. Forecasting, 1, 23–41, https://doi.org/10.1175/1520-0434(1986)001<0023:CMOPS>2.0.CO;2.
Bui, H., and T. Spengler, 2021: On the influence of sea surface temperature distributions on the development of extratropical cyclones. J. Atmos. Sci., 78, 1173–1188, https://doi.org/10.1175/JAS-D-20-0137.1.
Crespo, J. A., D. J. Posselt, and S. Asharaf, 2019: CYGNSS surface heat flux product development. Remote Sens., 11, 2294, https://doi.org/10.3390/rs11192294.
Crespo, J. A., C. M. Naud, and D. J. Posselt, 2021: CYGNSS observations and analysis of low-latitude extratropical cyclones. J. Appl. Meteor. Climatol., 60, 527–541, https://doi.org/10.1175/JAMC-D-20-0190.1.
Dacre, H. F., O. Martinez-Alavarado, and C. O. Mbengue, 2019: Linking atmospheric rivers and warm conveyor belt airflows. J. Hydrometeor., 20, 1183–1196, https://doi.org/10.1175/JHM-D-18-0175.1.
Demirdjian, R., J. D. Doyle, P. M. Finocchio, and C. A. Reynolds, 2022: On the influence of surface latent heat fluxes on idealized extratropical cyclones. J. Atmos. Sci., 79, 2229–2242, https://doi.org/10.1175/JAS-D-22-0035.1.
Edson, J., and Coauthors, 2013: On the exchange of momentum over the open ocean. J. Phys. Oceanogr., 43, 1589–1610, https://doi.org/10.1175/JPO-D-12-0173.1.
Field, P. R., and R. Wood, 2007: Precipitation and cloud structure in midlatitude cyclones. J. Climate, 20, 233–254, https://doi.org/10.1175/JCLI3998.1.
Gelaro, R., and Coauthors, 2017: The Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). J. Climate, 30, 5419–5454, https://doi.org/10.1175/JCLI-D-16-0758.1.
Grandey, B. S., P. Stier, R. G. Grainger, and T. M. Wagner, 2013: The contribution of the strength and structure of extratropical cyclones to observed cloud–aerosol relationships. Atmos. Chem. Phys., 13, 10 689–10 701, https://doi.org/10.5194/acp-13-10689-2013.
Gutowski, W. J., Jr., and W. Jiang, 1998: Surface-flux regulation of the coupling between cumulus convection and baroclinic waves. J. Atmos. Sci., 55, 940–953, https://doi.org/10.1175/1520-0469(1998)055<0940:SFROTC>2.0.CO;2.
Hartmann, D. L., 2016: Global Physical Climatology. 2nd ed. Elsevier, 473 pp.
Haualand, K. F., and T. Spengler, 2020: Direct and indirect effects of surface fluxes on moist baroclinic development in an idealized framework. J. Atmos. Sci., 77, 3211–3225, https://doi.org/10.1175/JAS-D-19-0328.1.
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, https://doi.org/10.1029/2012GL053866.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Hewson, T. D., 1998: Objective fronts. Meteor. Appl., 5, 37–65, https://doi.org/10.1017/S1350482798000553.
Hirata, H., R. Kawamura, M. Nonaka, and K. Tsuboki, 2019: Significant impact of heat supply from the Gulf Stream on a “Superbomb” cyclone in January 2018. Geophys. Res. Lett., 46, 7718–7725, https://doi.org/10.1029/2019GL082995.
Holton, J. R., 1992: An Introduction to Dynamic Meteorology. Academic Press, 511 pp.
Hoskins, B. J., and K. I. Hodges, 2002: New perspectives on the Northern Hemisphere winter storm tracks. J. Atmos. Sci., 59, 1041–1061, https://doi.org/10.1175/1520-0469(2002)059<1041:NPOTNH>2.0.CO;2.
Hoskins, B. J., and K. I. Hodges, 2005: A new perspective on Southern Hemisphere storm tracks. J. Climate, 18, 4108–4129, https://doi.org/10.1175/JCLI3570.1.
Houze, R. A., Jr., 1993: Cloud Dynamics. Academic Press, 570 pp.
Huffman, G. J., and Coauthors, 2017: NASA Global Precipitation Measurement (GPM) Integrated Multi-satellitE Retrievals for GPM (IMERG). Algorithm Theoretical Basis Doc., version 4.6, 28 pp., https://pmm.nasa.gov/sites/default/files/document_files/IMERG_ATBD_V4.6.pdf.
Huffman, G. J., E. F. Stocker, D. T. Bolvin, and E. J. Nelkin, and J. Tan, 2019: GPM IMERG final precipitation L3 half hourly 0.1 degree × 0.1 degree v06. Goddard Earth Sciences Data and Information Services Center, accessed 10 January 2022, https://doi.org/10.5067/GPM/IMERG/3B-HH/06.
Kuo, Y.-H., S. Low-Nam, and R. J. Reed, 1991: Effects of surface energy fluxes during the early development and rapid intensification stages of seven explosive cyclones in the western Atlantic. Mon. Wea. Rev., 119, 457–476, https://doi.org/10.1175/1520-0493(1991)119<0457:EOSEFD>2.0.CO;2.
Li, X., J. Yang, Y. Yang, and W. Li, 2022: Exploring CYGNSS mission for surface heat flux estimates and analysis over tropical oceans. Front. Mar. Sci., 9, 1001491, https://doi.org/10.3389/fmars.2022.1001491.
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, https://doi.org/10.1175/JCLI-D-12-00720.1.
Mak, M., 1998: Influence of surface sensible heat flux on incipient marine cyclogenesis. J. Atmos. Sci., 55, 820–834, https://doi.org/10.1175/1520-0469(1998)055<0820:IOSSHF>2.0.CO;2.
Naud, C. M., and B. M. Kahn, 2015: Thermodynamic phase and ice cloud properties in Northern Hemisphere winter extratropical cyclones observed by Aqua AIRS. J. Appl. Meteor. Climatol., 54, 2283–2303, https://doi.org/10.1175/JAMC-D-15-0045.1.
Naud, C. M., A. D. Del Genio, and M. Bauer, 2006: Observational constraints on the cloud thermodynamic phase in midlatitude storms. J. Climate, 19, 5273–5288, https://doi.org/10.1175/JCLI3919.1.
Naud, C. M., A. D. Del Genio, M. Bauer, and W. Kovari, 2010: Cloud vertical distribution across warm and cold fronts in CloudSat–CALIPSO data and a general circulation model. J. Climate, 23, 3397–3415, https://doi.org/10.1175/2010JCLI3282.1.
Naud, C. M., J. F. Booth, D. J. Posselt, and S. C. van den Heever, 2013: Multiple satellite observations of cloud cover in extratropical cyclones. J. Geophys. Res. Atmos., 118, 9982–9996, https://doi.org/10.1002/jgrd.50718.
Naud, C. M., J. F. Booth, and A. D. Del Genio, 2016: The relationship between boundary layer stability and cloud cover in the post-cold-frontal region. J. Climate, 29, 8129–8149, https://doi.org/10.1175/JCLI-D-15-0700.1.
Naud, C. M., J. F. Booth, M. Lebsock, and M. Grecu, 2018: Observational constraint for precipitation in extratropical cyclones: Sensitivity to data sources. J. Appl. Meteor. Climatol., 57, 991–1009, https://doi.org/10.1175/JAMC-D-17-0289.1.
Naud, C. M., J. Jeyaratnam, J. F. Booth, M. Zhao, and A. Gettelman, 2020: Evaluation of modeled precipitation in oceanic extratropical cyclones using IMERG. J. Climate, 33, 95–113, https://doi.org/10.1175/JCLI-D-19-0369.1.
Naud, C. M., J. A. Crespo, and D. J. Posselt, 2021: On the relationship between CYGNSS surface heat fluxes and the life cyclone of low-latitude ocean extratropical cyclones. J. Appl. Meteor. Climatol., 60, 1575–1590, https://doi.org/10.1175/JAMC-D-21-0074.1.
Otkin, J. A., and T. J. Greenwald, 2008: Comparison of WRF model-simulated and MODIS-derived cloud data. Mon. Wea. Rev., 136, 1957–1970, https://doi.org/10.1175/2007MWR2293.1.
Papritz, L., F. Aemisegger, and H. Wernli, 2021: Sources and transport pathways of precipitating waters in cold season deep north Atlantic cyclones. J. Atmos. Sci., 78, 3349–3368, https://doi.org/10.1175/JAS-D-21-0105.1.
Platnick, S., P. Hubanks, K. Meyer, and M. D. King, 2015: MODIS atmosphere L3 monthly product (08_L3). NASA Goddard Space Flight Center, accessed 20 March 2022, https://doi.org/10.5067/MODIS/MYD08_M3.006.
Rudeva, I., and S. K. Gulev, 2011: Composite analysis of North Atlantic extratropical cyclones in NCEP–NCAR reanalysis data. Mon. Wea. Rev., 139, 1419–1446, https://doi.org/10.1175/2010MWR3294.1.
Ruf, C. S., S. Asharaf, R. Balasubramaniam, S. Gleason, T. Lang, D. McKague, D. Twigg, and D. Waliser, 2019: In-orbit performance of the constellation of CYGNSS hurricane satellites. Bull. Amer. Meteor. Soc., 100, 2009–2023, https://doi.org/10.1175/BAMS-D-18-0337.1.
Salomonson, V. V., W. L. Barnes, P. W. Maymon, H. E. Montgomery, and H. Ostrow, 1989: MODIS: Advanced facility instrument for studies of the Earth as a system. IEEE Trans. Geosci. Remote Sens., 27, 145–153, https://doi.org/10.1109/36.20292.
Simmonds, I., K. Keay, and J. A. T. Bye, 2012: Identification and climatology of Southern Hemisphere mobile fronts in a modern reanalysis. J. Climate, 25, 1945–1962, https://doi.org/10.1175/JCLI-D-11-00100.1.
Tselioudis, G., Y. Zhang, and W. R. Rossow, 2000: Cloud and radiation variations associated with northern midlatitude low and high sea level pressure regimes. J. Climate, 13, 312–327, https://doi.org/10.1175/1520-0442(2000)013<0312:CARVAW>2.0.CO;2.
Tsopouridis, L., C. Spensberger, and T. Spengler, 2021: Characteristics of cyclones following different pathways in the Gulf Stream region. Quart J. Roy. Meteor. Soc., 147, 392–407, https://doi.org/10.1002/qj.3924.
