1. Introduction
Energy supplied by underlying warm ocean water is fundamental to the development of deep atmospheric convection of tropical cyclones and at times is possibly a pivotal factor in the maintenance and intensification of hurricanes [tropical cyclones that attain wind speeds greater than 33 m s−1 (64 kt or 74 mph)]. Yet, there are many oceanic and atmospheric processes controlling ocean surface temperatures, before and during a tropical cyclone event, that if better understood might help advance our ability to forecast the evolution of tropical cyclones, or more specifically, the likelihood of hurricane intensification. For example, in the Atlantic Ocean upper-ocean warm core rings are thought to have played a role in hurricane intensification (Goni and Trinanes 2003; Bosart et al. 2000; Shay et al. 2000). The impact of sea surface temperature (SST) and upper-ocean heat content on tropical cyclones continues to be studied from many perspectives (Free et al. 2004; Goni et al. 2004; Perrie et al. 2004; Shay et al. 2004; Lin et al. 2003). Underscoring the potential importance of the processes controlling ocean surface temperatures during a tropical cyclone event, a statistical study of 23 Atlantic hurricanes correlated the change of the SST from before an event to during an event, to the change in hurricane intensity (Cione and Uhlhorn 2003).
In this paper, two distinct ocean processes are presented from the perspective of how they define the early southernmost oceanic conditions experienced by more than half of all Atlantic hurricanes just prior to reaching the Caribbean Sea: 1) the spreading of freshwater discharges from the Orinoco and Amazon Rivers out into the western equatorial Atlantic Ocean, and 2) the periodic movement of North Brazil Current (NBC) rings through the river plumes. At the sea surface the freshwater discharges from the Amazon and Orinoco Rivers spread outward, forming extensive low-salinity plumes. The ∼0.2 × 106 m3 s−1 discharge from the Amazon River is the world’s largest and the ∼0.03 × 106 m3 s−1 discharge from the Orinoco River is the world’s third largest (Dagg et al. 2004; Perry et al. 1996). Pailler et al. (1999) map the freshwater barrier layer of the Amazon and Orinoco River plumes that can trap heat in the surface layer (0–30 m) of the ocean, resulting in warm SST anomalies, especially in August through October, and they discuss associated ocean–atmosphere exchange climate implications. In the same region, Ffield (2005) describes the SST signatures of anticyclonic NBC rings as they pass northwestward through the Amazon–Orinoco River plume with relatively cool SSTs during July through December in comparison to the warm SSTs associated with the plume.
The Amazon–Orinoco River plume and NBC rings are assessed in this paper specifically during the 1 June through 30 November Atlantic hurricane season, to identify their impact on the upper-ocean temperatures in the region, as well as to determine their proximity to passing hurricanes. The goal of this study is to provide the foundation for future oceanic boundary layer hurricane modeling studies, as well as for more systematic observational studies. Ultimately, the goal is to determine if the Amazon–Orinoco River plume and NBC rings are active participants, rather than passive bystanders, to the evolution of hurricanes, either by contributing to their maintenance or by impacting changes in their intensity.
2. Data
Historical in-situ ocean temperature and salinity profile data are provided by the National Oceanographic Data Center (NODC, released in 2002).
“Atlantic Tropical Storm and Hurricane Tracks” data are from the National Hurricane Center (http://www.nhc.noaa.gov/). The dataset includes tropical storms [tropical cyclones that attain wind speeds greater than 17 m s−1 (34 kt or 39 mph)] and hurricanes [as noted above, tropical cyclones that attain wind speeds greater than 33 m s−1 (64 kt or 74 mph)], but does not include tropical depression data [tropical cyclones with wind speeds less than 17 m s−1 (34 kt or 39 mph)].
Microwave satellite–derived SST measurements are from the Tropical Rainfall Measuring Mission (TRMM) Microwave Imager (TMI) beginning in early December 1997. TMI data are produced by Remote Sensing Systems and sponsored by the National Aeronautics and Space Administration (NASA) Earth Science Research, Education and Applications Solution Network (REASoN) Project (data are available online at www.remss.com). The advantage of the TRMM Microwave Imager (Wentz and Meissner 2000) is its ability to penetrate clouds, whereas the more traditional infrared-based satellite SST measurements are blocked by cloud cover. In general the 0.25° × 0.25° gridded TRMM daily dataset has at least one measurement listed every 3 days at each grid point, and often there are two measurements listed per day at each grid point, although the true resolution is actually less than this. Satellite sampling limitations of the TRMM data are primarily associated with the aliasing of the diurnal solar heating cycle. While there is typically an SST measurement listed at least once every 3 days in the TRMM dataset at each grid point, it actually takes about 46.7 days to repeat an SST measurement at approximately the same time of day due to the orbit of the satellite. Therefore, the diurnal SST cycle is manifested within the SST time series as a 46.7-day signal. In this region the highest aliasing (and therefore the largest diurnal cycles) is contained along the coast (Ffield 2005).
Infrared satellite–derived SST measurements are from the 7-day averaged Reynolds optimal interpolation SST dataset beginning in November 1981 and provided by the National Oceanic and Atmospheric Administration–Cooperative Institute for Research in Environmental Sciences (NOAA–CIRES) Climate Diagnostics Center (Boulder, Colorado; www.cdc.noaa.gov).
In general, in situ “SSTs” may be cooler than the satellite-derived SSTs, as a result of the in situ data incorporating the top meter or so of the ocean, whereas the satellite data measures the top centimeter or so of the ocean.
3. Background and analysis
a. Amazon and Orinoco River plumes
The freshwater discharges from the Amazon and Orinoco Rivers spread outward into the western equatorial Atlantic Ocean while continually mixing with surrounding salty ocean surface water. The geographical distribution of the low-salinity signatures of the Amazon and Orinoco River plumes can be revealed with historical in situ surface salinity data. The low salinity from the Amazon River freshwater plume flows to the northwest along the coast (Ou 1989) from January to June (Ffield 2005) until it reaches the eastern edge of the Guiana Plateau (52°W), where it spreads broadly northward. From July to December (Ffield 2005) the Amazon River freshwater signal is also observed to flow to the northwest along the coast, but then at the eastern edge of the Guiana Plateau it extends northward, and then both westward (merging with the Orinoco River plume) and eastward (Lentz and Limeburner 1995; Lentz 1995).
The spreading of the Amazon and Orinoco River plumes into the western equatorial Atlantic Ocean is the topic of numerous papers (e.g., Gonzalez-Silvera et al. 2004; Astor et al. 2003; Hellweger and Gordon 2002; Kelly et al. 2000; Signorini et al. 1999; Muller-Karger et al. 1988), and the plumes are pertinent to a host of chemistry (Koertzinger 2003; Cutter et al. 2001; Ternon et al. 2000), biology (Carpenter et al. 2004), fisheries (de Moura et al. 2001), and geology (Warne et al. 2002) studies. The SST in the region is linked to Amazon Basin rainfall (Fu et al. 2004; Labat et al. 2004; Ronchail et al. 2002; Wang and Fu 2002; Fu et al. 2001; Liebmann and Marengo 2001) and the intertropical convergence zone (Chang et al. 2000). In addition, the plumes spread into the region classified as the tropical Western Hemisphere warm pool, defined by SSTs greater than 28.5°C, and associated with increased tropical convection (Wang and Enfield 2001). The annual cycle of tropical Western Hemisphere warm pool intensity and areal extent is associated with the annual development of tropical storms and hurricanes (Wang and Enfield 2001).
1) Salinity signature of the plume
Coinciding with the Atlantic hurricane season, the plumes from both the Amazon and Orinoco Rivers combine and spread to their average maximum historical extent during 1 June through 30 November, as revealed by the 35.4 salinity contour of the averaged historical surface salinity data (Fig. 1a) reaching northward (22°N, 58°W), westward (15°N, 77°W), and eastward (7°N, 34°W). The average surface salinity inside this 35.4 salinity contour is less than 35.4 and is freshest closest to the river outlets. The 35.4 salinity contour is subsequently used in this study to define the historical average extent of the Amazon–Orinoco River plume during hurricane season.
The Amazon–Orinoco River plume is farthest eastward simultaneous with the height of the hurricane season from mid-August through mid-October. The maximum monthly Amazon River discharge is in May and June (Hellweger and Gordon 2002). The maximum monthly Orinoco River discharge is in August (two to three months after the Amazon; note that the Orinoco’s maximum flow is less than the seasonal minimum flow of the Amazon) (Hellweger and Gordon 2002). In the eastern extension of the plume (7°N, 34°W), the seasonal cycle of the river-influenced, low-salinity ocean surface waters is from July to October, with a minimum salinity of 32.8 (an extremely low open-ocean value) in August, which is two to three months after the maximum in the Amazon River discharge and coincides with the height of the hurricane season; during the rest of the year, November to June, the salinity averages 36.0 in that region (a typical open-ocean value), indicating that there is no freshwater river discharge influence at that location during those months (Ffield 2005). For the westward extension of the plume (15°N, 77°W), low salinities are observed entering the Caribbean, and a high inverse correlation (R2 = 0.92) is found between the annual cycle of ocean surface salinity near Barbados (approximately located at 13°N, 59°W) and the 2-month lagged Amazon River discharge, which is concluded to predominate over the Orinoco River discharge (Hellweger and Gordon 2002). The ocean surface salinity near Barbados is minimum in June (34.24), the beginning of the Atlantic hurricane season, and maximum in January (35.48), two months after hurricane season is finished (Hellweger and Gordon 2002).
The large size of the plume is the result of the tremendous freshwater outflows from the Amazon and Orinoco Rivers that discharge almost 20% of the total annual global freshwater river outflow directly to the surface waters of the western equatorial Atlantic and Caribbean. While the Amazon River influence is probably most important—its discharge is nearly 10 times larger than the Orinoco’s discharge—in this study both rivers are generally referred to because signatures from each river are clearly observed just offshore of their respective deltas, but it is beyond the scope of this paper to distinguish between the two signals farther offshore where they appear merged. The combined freshwater river discharge can affect a considerable ocean volume, because, as in this case, the regional ocean with salinities generally greater than 36 can be diluted by just a tiny portion of zero-salinity river water to form modified plume waters with a dynamically significant anomalous salinity signature, in this case 35.4 salinity. Or considered in a different way, 0.23 Sv [1 Sverdrup (Sv) ≡ 1 × 106 m3 s−1] of Amazon and Orinoco freshwater river discharges mixed with 36 salinity ocean water can “produce” 13.6 Sv “flow” of 35.4 salinity plume water, a volume transport comparable to the upper 50-m-layer flows of the ocean’s major western boundary currents.
2) Temperature signature of the plume
The Amazon and Orinoco River plumes are potentially important to tropical cyclone evolution, because the plume is associated with the warmest temperatures in the region (Pailler et al. 1999; Ffield 2005). Pailler et al. (1999) reveal the dynamics of the temperature association by mapping the barrier layer connected with the freshwater discharge of the Amazon River as it spreads into the western equatorial Atlantic Ocean, inhibiting mixing between surface and subsurface waters and trapping solar heating within the 3–30-m ocean surface layer due to the sharp vertical contrast in density. Here, the temperature signal associated with the low-salinity waters of the plume is isolated by averaging the surface temperatures of only stations with surface salinities less than 33.5 (a relatively restrictive cutoff because the average plume extent was defined above using the saltier 35.4 salinity contour). This restricts nearly all the remaining data to the region within the plume boundaries (as defined by the 35.4 salinity contour), and the core of warmest temperatures associated with the plume is revealed by the 28.0°C contour (Fig. 1b) reaching northward (15°N, 64°W), westward (14°N, 66°W), and eastward (7°N, 34°W) during the June through November hurricane season months. Inside the 28.0°C temperature contour, average plume surface temperatures greater than 28.5°C are revealed in the eastern extension of the plume. The mapping of the river plume barrier layer (Pailler et al. 1999) is similar to this mapping of the core of warmest temperatures associated with the plume. (Note that even higher SSTs are observed in the Caribbean Sea, which is not included in this study.)
In comparison to the warm temperatures associated with the plume, the ocean surface temperatures to the north and to the northeast are much cooler, averaging between 25° and 27°C during hurricane season (Fig. 1c). By extracting only the extreme temperature values during hurricane season from the dataset, the locations of the potentially warmest versus coolest regions are exposed: ocean surface temperatures warmer than 29°C are found in the plume region versus cooler than 26°C both to the northeast and to the southeast (in the NBC), a contrast of more than 3°C (Fig. 1d).
3) Buoyancy signature of the plume
The plume’s pool of buoyancy contributes to its persistence to remain as a distinct surface layer and inhibits the potential for deep-reaching wind mixing due to its high stability. The average salinity profile for the eastern extension of the plume (not shown) reveals the low-salinity signature reaching to 50-m depth in contrast to the much higher salinities in the upper 50 m in the neighboring regions. Viewed in terms of density (combining the buoyancy effects of temperature and salinity; not shown), even in the far offshore northern section of the plume the density in the surface (22.3 sigma theta) and in the upper 50 m (22.9 sigma theta) is far more buoyant (and therefore stable) than the surrounding waters (surface and upper 50-m densities all 23.7 sigma theta and denser). During hurricane season the plume varies from 10- to 60-m depth, with an average of 33 m, defined by restricting salinities to less than 35.4 at the sea surface and less than 35.8 at depth; the average minimum (maximum) plume depths are 10 to 20 m (20 to 60 m) deep. However, it is likely that the plume is patchy at times, both in ocean–atmosphere area contact and in depth, varying from thin patches of insignificant ∼2- to 5-m slicks, to a continuous, extensive ∼33-m-thick plume. Therefore, seasonal and interannual variability in the temperature, stability, area, and depth of the plume are probably important factors to consider, but all are outside the scope of this paper.
4) Historical tropical cyclones tracks
Examples of individual hurricane tracks reveal that hurricanes often pass directly through the northern section of the Amazon–Orinoco River plume. In Fig. 2, the tracks for tropical cyclones that at some point attain maximum wind speeds of 83 kt or greater (i.e., category 2 hurricanes and greater) for the 1960 to 1965 time period are shown, as an example, revealing that 14 of the hurricanes pass over the historical region of the plume (yellow lines) and 12 do not (thin blue lines). In fact, as hurricane strength increases, an increasing percentage of hurricanes pass through the plume region: for the 1960 to 2000 time period 68% of the 19 hurricanes that attained maximum wind speeds greater than 135 kt (i.e., category 5 hurricanes) passed directly over the historical region of the plume, whereas only 32% never passed over the plume (Table 1). This reveals that most of the most destructive hurricanes may be directly influenced by ocean–atmosphere interaction with the plume just prior to reaching the Caribbean. Tallied within subregions, of all 1960 to 2000 hurricanes, 16% of those passing over the warm plume ultimately attain category 5 strength, whereas only 4% of those passing outside of the plume ultimately attain category 5 strength (Table 2). In contrast, the weakest hurricanes (category 1) typically do not pass over the plume (only 17%), and instead most (83%) miss the warm plume temperatures and pass outside (i.e., north) of the plume (Table 1). Hurricanes are cyclonic and typically have 300- to 1000-km diameters; therefore the ocean–atmosphere contact between the hurricanes and the plume is much larger than the one-dimensional hurricane track lines convey. The southernmost hurricanes, those with tracks south of 14°N, are likely to be most directly influenced by the warmest core temperatures of the plume.
Using all the hurricane track data reveals that Atlantic tropical cyclones historically pass just north of the core of the warmest surface temperatures of the Amazon–Orinoco River plume and directly over the northernmost portion of the low-salinity reach of the plume. By binning all tropical storm and hurricane tracks from 1950 through 2003 in one-degree squares, the historical tropical cyclone corridor is revealed along approximately 12° to 20°N (Fig. 3). As suggested above by the hurricane track examples (Fig. 2), this corridor passes through the northernmost portion of the average historical position of the Amazon Plume defined by the 35.4 salinity contour (Fig. 3, green contour) and associated with 2°C on average warmer surface temperatures (Fig. 4, red circle symbols). The corridor also passes alongside, just north of, the >28°C maximum temperatures associated with the plume for a distance of about 3300 km (Fig. 3, red contour).
5) SSTs encountered by hurricanes
As hurricanes move westward across the Atlantic in the 12° to 20°N latitude band they abruptly encounter much warmer ocean water when they reach the Amazon–Orinoco River plume. Warm SSTs in the northernmost reach of the plume (22°N, 58°W), north of the core of warmest plume temperatures, are also linked to the freshwater discharge of the rivers, but a less restrictive salinity cutoff must be used in order to allow for the reduced intensity of the freshwater river signal with distance from shore due to vertical mixing with subsurface saltier waters as well as horizontal mixing with regional saltier surface waters (evaporation may also be a factor). Pailler et al. (1999) show this temperature association with a temperature–salinity diagram. In a similar fashion, by contrasting stations with surface salinities less than 35.8 versus those with surface salinities greater than 35.8 during the June through November hurricane season, the temperature–salinity scatterplot for data within the 12° to 20°N latitude band independently reveals the warm temperature signal and geographical position of the Amazon–Orinoco River plume at these more northern latitudes (Fig. 4). In this case, the average surface temperatures first encountered by tropical cyclones moving westward between 12° and 20°N is only 26°C (Fig. 4, blue symbols), but upon reaching the Amazon–Orinoco River plume, identified by the lower salinities, the average surface temperature encountered by tropical cyclones is 2°C warmer, 28°C (Fig. 4, red symbols). Seasonal and interannual variability in the northward spreading of the Amazon–Orinoco River plume may also be an important factor controlling the ocean temperatures encountered by hurricanes in the 12° to 20°N latitude band at this location.
6) Upper-ocean temperatures and salinities encountered by hurricanes
In outlining the SST conditions during hurricane season, the ocean characteristics below the air–sea boundary also need to be considered, because warm SSTs during a tropical cyclone event can be cooled quickly by vertical mixing with colder subsurface water. In general the subsurface temperatures mirror the SST trends, such that regions with warmer (cooler) SSTs also have relatively warmer (cooler) subsurface temperatures.
In the zonal strip that is frequently traversed by tropical cyclones (same zonal strip as in Fig. 4), the northern plume temperatures are consistently very warm during hurricane season throughout the upper ocean, averaging from 28°C at the sea surface down to 25.5°C at 100-m (∼db) depth, with a narrow standard deviation envelope (Fig. 5a, red). These warm temperatures are again associated with the river-influenced low-salinity signature, which can be easily discerned from the ocean surface down to 50 m with salinities less than 36 (Fig. 5b, red). In contrast, to the east in the strip frequently traversed by tropical cyclones, as well as to the north of the strip, the ocean temperatures throughout the upper ocean during hurricane season are much cooler, averaging between 25° and 26°C at the surface, down to around 22°C at 100 m, with a large standard deviation envelope (Fig. 5a, blue and green), and the salinities are all higher than 36 (Fig. 5b, blue and green). Consequently, even with deep-reaching mixing down to 100-m depth into the ocean, hurricanes passing over the plume region will abruptly encounter significantly warmer, and more stable, ocean temperatures.
b. NBC retroflection and rings
To fully assess the impact of the warm Amazon–Orinoco River plume on hurricanes, or even to just properly interpret satellite SST images in the region, the role of NBC rings that periodically move through the plume region must be considered. The NBC flows northwesterly along the coast of Brazil, retroflects near 6°N, 48°W, and then flows easterly; at the retroflection the NBC periodically sheds 300–500-km-diameter anticyclonic rings that translate toward the Caribbean Sea (Goes et al. 2005; Fratantoni and Richardson 2006; Fratantoni and Glickson 2002; Garraffo et al. 2003; Garzoli et al. 2003; Goni and Johns 2003; Johns et al. 2003; Wilson et al. 2002; Limeburner et al. 1995). Historical salinity data during hurricane season highlight the juxtaposition between the NBC retroflection, with high surface salinities, and the Amazon–Orinoco River plume, with low surface salinities (35.4 salinity contour; 6°N, 48°W; Fig. 1a). During hurricane season, the NBC rings carry relatively cool ocean surface temperatures from the NBC (down to less than 26°C) directly through the middle of the historical location of the warm Amazon–Orinoco River plume (up to more than 29°C; Fig. 1d). This is opposite to 1) the January to June time period when the SSTs of the rings and retroflection are relatively warm in comparison to the regional ocean SSTs (Ffield 2005), and 2) the relatively warm SSTs associated with NBC rings when they are in the Caribbean Sea (Goni et al. 2004; Goni and Trinanes 2003). The NBC rings are on the order of half the diameter of Atlantic hurricanes, and they are slow moving, taking a month or so to move through the plume region. Therefore, the up to 3°C cooler surface temperatures of the rings may cover a sufficiently large ocean surface area, for a long enough time, to diminish the warm temperature effect of the much larger and more visible pool of warm temperatures associated with the Amazon–Orinoco River plume.
It is of particular importance to note that it takes extra effort to identify the cool surface temperatures of NBC rings when using satellite SST data, because at the sea surface warm Amazon–Orinoco River plume waters can partially or completely hide their presence. Compounding the difficulty, there are only a few detailed oceanographic surveys of NBC rings, especially during hurricane season, which could help guide proper interpretation of satellite data. In the following example, Amazon–Orinoco River plume water is observed to completely cover the surface of an NBC ring. In December 1998 (Fig. 6a), a ring was surveyed (Wilson et al. 2002) that probably shed from the NBC retroflection around October 1998 during hurricane season, and therefore may be characteristic of rings during hurricane season. The ring displays warm (>28.4°C) and low-salinity, river-influenced (<35.4) plume water in the top-most 30–40 m, centered over the core of the spinning ring. At deeper levels the spinning ring carries the cooler and saltier South Atlantic water characteristic of the NBC. In this situation concurrent satellite SST observations will not reveal the presence of the cool near-surface NBC ring waters, because they are hidden just underneath the warm plume. [However, the ring can be successfully monitored by continuous tracking of the SST fields (“TMI-98-7”; Ffield 2005).] Therefore, the influence of the warm Amazon–Orinoco River plume on hurricanes cannot be solely determined by a quick view of satellite SST observations, as relatively cool NBC rings hiding just underneath the warm plume can contribute to much faster cooling of SST as hurricanes pass by.
Strong winds from the 300–1000-km-diameter cyclonic hurricanes might quickly obliterate a thin plume, exposing several degrees cooler NBC ring water directly to the atmosphere midstorm. Possibly there is a relationship dependent on the translation speed of hurricanes with slow and vigorous hurricanes most likely to mix through to the cool temperatures of rings. Alternatively, it is plausible that instead a slow hurricane may be able to intensify by continually drawing energy from a plume that is sufficiently deep and stable enough to shield the hurricane from any effect of cool ring waters hidden underneath.
A second example of an NBC ring is shown to point out how Amazon–Orinoco River plume water can also just partially hide the presence of a ring. In this case the plume water is entrained just at a ring’s edges, so concurrent satellite SST observations should detect most of the near-surface temperature signal associated with the ring. In February 1999 (Fig. 6b), a ring was surveyed (Wilson et al. 2002) that probably shed from the NBC retroflection around December 1998, just after hurricane season ended. The ring reveals high-salinity (>36.2) South Atlantic water throughout its upper layers, with low-salinity (<35.4) plume waters caught only in the top-most 30 m at the edges of the spinning ring. The ring carries temperatures around 27°C throughout the upper layers (which are actually warmer than the surrounding waters as this example is not during hurricane season). SSTs associated with this ring are easily observed in satellite SST observations (“TMI-99-1”; Ffield 2005), in contrast to the previous example, demonstrating the variability in the SST signature of NBC. Therefore, while the potentially cooling effect of NBC rings on the SSTs during hurricane season in the Amazon–Orinoco River plume region should be considered, it is not necessarily an easy task, because a thin layer of warm plume water can partially or completely hide the potentially more substantive cool near-surface temperatures of a ring.
c. Two hurricane case examples
Two hurricane case examples are shown to illustrate the hypothesis of the possible role of the Amazon–Orinoco River plume ocean region as a participant in the evolution of hurricanes. Microwave (cloud-penetrating) satellite-derived SST data are used in juxtaposition with wind-speed coded hurricane tracks. In both examples, the warm SSTs of the plume and the intruding cool SSTs of the NBC retroflection from the south are evident. However, the contrasting sizes and temperatures of the plume and the presence of an NBC ring differentiate the two cases, and these factors are presented as examples of the possible influences to consider in future, dedicated studies on the demise, in the first case, and the maintenance/intensification, in the second case, of hurricanes passing through the plume region.
1) Hurricane Joyce, 2000
The Amazon–Orinoco River plume SST conditions a few days prior to Hurricane Joyce were not particularly extensive or warm for the time of year, and the cool NBC retroflection had pushed relatively far northwest into the plume. In September–October of 2000, Hurricane Joyce became a hurricane after passing west of 38°W, attaining an 80-kt wind speed (category 1) at 12.2°N, 42.5°W, but then surprisingly dropped to lower wind speeds (−5 kt over 6 h) after passing west of 10.9°N, 46.1°W (Lawrence 2000; color-coded circles, Fig. 7a). Before the hurricane, the warmest plume SSTs were >29.5°C, and were just to the south of the imminent center track path of Hurricane Joyce (Fig. 7a). In the region bounded by 5°–15°N, 55°–35°W, only 40% of the SSTs were warmer than 29.0°C, but 13% were cooler than 28.0°C, quite cool conditions in comparison to the next case example. Also before the hurricane, the relatively cold (<28.5°C) NBC retroflection had pushed northwestward to 8°N, 50°W, about 220 km south of the future path of the center of the hurricane (in the southern part of the white box in Fig. 7a). From 8 yr of altimetry data, the average northerly position of the NBC retroflection is 6.7°N ± 1.8° (Fonseca et al. 2004). Therefore, the NBC retroflection just prior to Hurricane Joyce is more northerly than average.
Hurricanes passing through the western tropical Atlantic Ocean can expose cool NBC rings concealed just underneath the warm surface waters of the Amazon–Orinoco River plume or they can contribute to keeping the cool SSTs of a newly shed NBC ring exposed. The SST image a few days after Hurricane Joyce passed reveals that, between 52° and 48°W, the warm SSTs of the plume were obliterated and a broad meridional swath of cool SSTs was exposed (48°–53°W; Fig. 7b). In addition, there is now, with SSTs less than 28°C, a characteristic circular surface signature of a cool NBC ring isolated within the warm surface waters of the Amazon–Orinoco River plume (southern half of white box, Fig. 7b). The cool SST features coincide, whether coincidentally or causally, with the hurricane winds dropping suddenly from their maximum levels, and subsequently dissipating altogether just to the west. While directly over the swath of cool SSTs, the winds dropped from 70 to 55 kt. The maximum SST contrast between the warm plume (>29.5°C) and the ring feature (<27.5°C) is 2°C. The SST difference map, calculated as the difference between the “after” and “before” SST images of Hurricane Joyce, and expanded for the NBC ring generation region, clearly reveals the sharply delineated ring feature with SSTs −0.9° to −1.5°C cooler than before the hurricane (Fig. 8c).
Hurricane Joyce was a relatively slow moving hurricane in comparison to the next case example, and this may have enabled sufficient time for the cool SSTs of the NBC ring to be exposed to the surface by the hurricane-strength winds. The westward translation speed of Hurricane Joyce averaged around 5 degrees of longitude per day (Fig. 8a). In addition, the relative coolness, and therefore probably weak stability, of the plume just prior to the hurricane passing may also have contributed to the potential ease in exposing the NBC ring to the atmosphere. It is concluded that the presence of a large, cool NBC ring, suddenly exposed within the warm plume waters during Hurricane Joyce, is an important factor to consider in studies of the hurricane’s demise.
2) Hurricane Ivan, 2004
The Amazon–Orinoco River plume SST conditions a few days before Hurricane Ivan were quite extensive and warm, and the NBC retroflection was also warmer. In September 2004, Hurricane Ivan, ultimately a category 5 hurricane and the strongest hurricane to be recorded so far south, continually intensified as it passed through the plume attaining a 115-kt wind speed (category 4) at 10.6°N, 48.5°W, at which point its wind speeds dropped (−2.5 kt over 6 h), briefly decreasing to 90 kt at 11.3°N, 54.4°W, from whence they began to increase again (Stewart 2005; Fig. 9a). Before the hurricane, the warmest SSTs (>29.5°C) in the core of the plume were directly in the center of the future path of Hurricane Ivan (Fig. 9a). The size of the region with temperatures >28.5°C was about twice as large (Fig. 9a) as it had been for Hurricane Joyce (Fig. 7a). In the region bounded by 5°–15°N, 55°–35°W, 78% of the SSTs were warmer than 29.0°C and none were cooler than 28.0°C just before Hurricane Ivan, a much warmer regime than it had been for Hurricane Joyce. Before the hurricane, the NBC retroflection is also at about 8°N, 50°W (southern part of the white box, Fig. 9a). However in this case the NBC retroflection is significantly warmer, with its northernmost tip 1.0°C warmer (29.5°C) than it was for Hurricane Joyce (28.5°C). SSTs cooler than 28.5°C in the NBC retroflection are more than 660 km away from the future center track of Hurricane Ivan, outside of its direct ocean–atmosphere contact, whereas they were only 220 km away, most likely in direct ocean–atmosphere contact of Hurricane Joyce.
Depending on seasonal and interannual variability, the Amazon–Orinoco River plume may be warm, deep, stable, and extensive enough such that the plume and its associated warm ocean surface temperatures are sustained as a hurricane passes. After Hurricane Ivan passed, the SSTs in the region indicate that despite the hurricane-force winds, the Amazon–Orinoco River plume retained its size and characteristically warm temperatures, with SSTs greater than 29.5°C (Fig. 9b). The warm plume SST features coincide, whether coincidentally or causally, with the hurricane winds intensifying over the plume. While directly over the 30°C (Fig. 9a) core of warmest plume SSTs, the winds increased from 50 to 110 kt. The SST difference map, calculated as the difference between the after and before SST images of Hurricane Ivan, and expanded for the NBC ring generation region, reveals unstructured, yet overall cool temperatures with SSTs down to −1.0°C cooler than before the hurricane (Fig. 8d). This leads to the conclusion that there is no evidence of an NBC ring impacting ocean–atmosphere interaction during the course of Hurricane Ivan, as there was for Hurricane Joyce.
Hurricane Ivan was a relatively fast moving hurricane, and this may have prohibited sufficient time for cool ocean temperatures underneath the warm Amazon–Orinoco River plume, if there were any, to ever be exposed to the atmosphere. The westward translation speed of Hurricane Ivan averaged around 7 degrees of longitude per day (Fig. 8b). In addition, the extreme warmness, and therefore probably strong stability, of the plume just prior to the hurricane passing may also have contributed to the persistence of the warm plume SSTs despite strong hurricane-force winds. It is concluded that the presence of a large, warm plume directly under the path of Hurricane Ivan for five days is an important factor to consider in future studies on the maintenance/intensification of Hurricane Ivan.
d. SST and tropical cyclone wind speed statistics
Creating hurricane season time series with data just from June through November months for each year reveals the known ENSO variability associated with both hurricanes and regional SSTs. All tropical cyclone wind speed values falling within the 7.9°–25.0°N, 55.0°–35.0°W block are averaged together for each June through November hurricane season, producing a single time series with a single average wind speed value for each year (Fig. 10a, “+” symbols). For temperature, infrared satellite-derived SSTs are also averaged for each June through November hurricane season, producing a time series at each one-degree square with a single average SST value for each year in 1988 through 2003. A single example of an SST time series from within the Plume is shown in Fig. 10a (“○” symbols). Both time series reveal the known association of hurricanes and regional SSTs with ENSO (Fig. 10b), such that, in this rendition, stronger-than-average tropical cyclone wind speeds and warmer-than-usual SSTs are observed during La Niña years (e.g., 1988, 1996, and 1999) and weaker-than-average wind speeds and cooler-than-usual SSTs during El Niño (e.g., 1987 and 1997).
A more general assessment of the association between tropical cyclone wind speeds and SSTs in the region of the Amazon–Orinoco River plume also isolates and identifies the plume as the most characteristic feature defining the ocean surface conditions in the region. The cross correlations between the hurricane season time series of tropical cyclone wind speed (Fig. 10a) and SST at each one-degree square are calculated. The cross-correlation coefficients (Fig. 11) reveal that high correlations (>0.50) are all generally within the historical position of the plume (as defined previously by the average in situ 28.0°C surface temperatures; Fig. 1b), and the highest correlations (>0.60) correspond to the two locations associated with the freshest plume waters stemming from the Orinoco River and the Amazon River mouths (Fig. 1a, two separate 34.7 salinity contours). In contrast, the correlations are quite low (<0.25) to the north of the plume even though the wind speed data from that region are included in the average wind speed time series (northern half of the large “square” on Fig. 11). The region for the wind speed time series was chosen to simply encompass the corridor of historical tropical cyclones. Consequently, the geographic pattern of cross correlations with SST independently, and with remarkably similar boundaries, highlights the importance of the SST in the region of the historical Amazon–Orinoco River plume. In addition, the high correlations in the plume region imply that warm plume SSTs may directly contribute to stronger hurricane wind speeds. (A causal linkage cannot be proven with this type of analysis.) La Niña is associated with increased rainfall over the nearby continent, which through river discharge is likely to correlate to a warmer (and/or more extensive) plume in the same year. La Niña is also associated with a more vigorous hurricane season. Therefore, warmer (cooler) plume SSTs due to increased (decreased) river discharge during La Niña (El Niño) may directly contribute to a more (less) vigorous hurricane season.
The highest cross correlation between the hurricane season time series of tropical cyclone wind speed and SSTs locates the center of the Amazon low-salinity plume water, which also happens to be the within the core of the northwestward translation pathway of NBC rings just after they shed from the NBC retroflection. The highest cross-correlation coefficient (0.67) is at 7.5°N, 52.5°W (Fig. 11), where the lowest sea surface salinities stemming from the Amazon River are observed as noted before (Fig. 1a, rightmost 34.7 contour). The hurricane season SST time series for this location was shown as the SST example in Fig. 10a (“○” symbols). The location is at the northern edge of the Guiana Plateau, a prominent feature reaching out from the coastline. The Guiana Plateau is commonly mentioned in connection with the shedding of NBC rings, and this connection is substantiated by all NBC ring drifter tracks for 10 rings clearly passing through the location (Fratantoni and Richardson 2006).
More in-depth analysis of the tropical cyclone wind speeds and SST at the northern edge of the Guiana Plateau (7.5°N, 52.5°W) suggests that 53-day earlier SSTs, possibly associated with the shedding of NBC rings, are most significantly related to passing tropical cyclones farther to the north. In this calculation, the average wind speed for each individual tropical cyclone (whereas before all the cyclones for a year were averaged together) passing through the corridor of historically intensifying tropical cyclones (the 7.9°–25.0°N, 55.0°–35.0°W block) is determined for the years 1982 through 2003. For each of the 52 resulting values, the infrared satellite-derived SST from 7.5°N, 52.5°W is identified at the same time as each tropical cyclone, and as a function of lagged time, for each tropical cyclone. Linear regressions (not shown) for all the lagged sets of 52 pairs of values are determined. The results (Fig. 12) reveal the most significant linear regression (r = 0.48) with SSTs that are 51.1 to 54.8 days earlier, such that—as with the annually averaged hurricane season time series—stronger (weaker) tropical cyclone wind speeds are associated with warmer (cooler) SSTs. (The second-most significant linear regression is with SSTs that are 14.6 days earlier, suggestive of fortnightly tides with a 14.76-day period, possibly linked to tidal currents responding to the Guiana Plateau.) The ∼53 days is suggestive of the role of NBC rings, because for a 1.6-yr observation period, the characteristic number of days between NBC retroflection shedding events was between 45 and 55 days, with an average rate of 49.4 days (Ffield 2005). In addition, a high-resolution general circulation model with variable winds shows that the North Equatorial Counter Current radiates Rossby waves with a period of 50 days that reflect at the Brazilian coast, producing NBC rings (Jochum and Malanotte-Rizzoli 2003). So perhaps because of the similar timing the processes are related in some (not obvious) way to the tropical cyclone wind speeds. Alternatively, observations from a current meter mooring indicate that the NBC rings take about 1 month for the leading and trailing edges to pass across a longitude with an average 12.5 km day−1 translation speed (Johns et al. 2003). Therefore, over 53 days a new NBC ring might move 634 km (equivalent to about 6°) northwestward into the more direct path of a westward-moving tropical cyclone, where the ring’s cool surface temperature could more directly impact the evolution of the tropical cyclone on its southern half. However by that time the SST signal of the NBC ring could be masked by a thin layer of warm plume water, possibly explaining why it would be the more southern, earlier, and unadulterated SST that shows the strongest linkage to the tropical cyclone wind speeds. For this reason, simply including the concurrent “skin-layer” SST field into hurricane prediction models might not be sufficient, as a hidden ring with a few degrees cooler temperatures, over an area on order of 71 000 to 196 000 km2 just below the ocean’s surface, with stabilities governed by both temperature and salinity, might be quickly reexposed with the passing of the hurricane.
Removing the interannual variability from the time series presents an additional perspective to view the data enabling a hypothesis on the impact of SSTs in the Amazon–Orinoco region on the evolution of passing tropical cyclones. When the 4-yr Gaussian running mean is removed from the average wind speed for each individual tropical cyclone passing through the corridor of historically intensifying tropical cyclones (the 7.9°–25.0°N, 55.0°–35.0°W block) and plotted as a function of the 4-yr anomaly 7.5°N, 52.5°W SSTs from 53 days earlier, it is revealed that earlier, warm SSTs at the northern edge of the Guiana Plateau are associated with both weak and strong tropical cyclone wind speeds, but earlier, cool SSTs—possibly from recently shed NBC rings—are almost always associated with weak tropical cyclone wind speeds and never associated with strong tropical cyclone wind speeds (Fig. 13). This result is not robust, as the plot using 51-day earlier SSTs (not shown) does not show as convincing a relationship (three points contradict). However, it does suggest the linkages that a more in-depth study might explore. Most importantly, a future study should consider the size and exact position of the tropical cyclones, distinguishing between cases with large (small) diameter tropical cyclones that may potentially have more (less) direct contact with the warm Amazon–Orinoco River plume and any cool-surface NBC rings in their path.
4. Conclusions
The freshwater discharges from the Amazon and Orinoco Rivers spread into the western equatorial Atlantic Ocean combining to form an extensive buoyant plume at the ocean surface with the lowest salinities and warmest temperatures in the region. The spreading of the Amazon–Orinoco River plume exhibits a seasonal cycle coinciding with the Atlantic hurricane season (1 June through 30 November) with river-influenced minimum salinities observed farthest eastward in August during the height of the hurricane season (mid-August to mid-October). For the 1960 to 2000 time period 68% of all category 5 hurricanes passed directly over the historical region of the plume, revealing that most of the most destructive hurricanes may be influenced by ocean–atmosphere interaction with the plume just prior to reaching the Caribbean. Average ocean surface temperatures first encountered by tropical cyclones moving westward between 12° and 20°N is only 26°C, but upon reaching the northern reaches of the Amazon–Orinoco River plume, identified by its river-influenced low salinities, the average ocean surface temperatures encountered by tropical cyclones are 2°C warmer: 28°C. (These results are from in situ data measuring the top few meters of the ocean; satellite-derived SST values are generally warmer as they measure the upper-most “skin” of the ocean.) On average, low salinities persist down to 80 m in the ocean and are associated with warmer temperatures up to +4°C at depth. As warm ocean surface temperatures may play a role in hurricane maintenance and intensification, understanding the processes and variability controlling the temperatures of the plume may be important. In addition, the buoyant, and therefore stable, 10- to 60-m-thick layer of the plume can mask other ocean processes and features just below the surface, in particular cool (during hurricane season) surface temperatures carried by NBC rings. After shedding from the NBC retroflection, the 300–500-km-diameter anticyclonic (clockwise) NBC rings pass northwestward through the Amazon–Orinoco River plume toward the Caribbean. The limited observations reveal that at times the cool upper-layer temperatures of the NBC rings are exposed to the atmosphere at the sea surface and at other times they are hidden just underneath warm plume water. Strong winds from the 300–1000-km-diameter cyclonic (counterclockwise) hurricanes might quickly obliterate a thin plume, exposing several-degrees-cooler NBC ring water to the surface, and potentially contributing to rapid cooling of SST midstorm. In this study, the warm temperatures associated with the low-salinity Amazon–Orinoco River plume and the relatively cool temperatures associated with NBC rings during hurricane season are shown to be in close proximity to hurricanes passing over the region. Their proximity draws attention to their possible role as active participants in hurricane maintenance and intensification. Any modeling study on the role of the oceanic boundary layer on hurricane maintenance and intensification in this region must consider the roles of the Amazon–Orinoco River plume and NBC rings on the upper-ocean temperature, depth, and stability (a function of temperature and salinity), as these processes are probably important and cannot be simply represented by including satellite-derived skin SSTs into models. In fact, the highest cross correlations between SSTs in the tropical Atlantic Ocean and average tropical cyclone wind speeds between 35° and 55°W are found only in the Amazon–Orinoco River plume region. This implies a possible causal relationship between warmer plume temperatures and stronger hurricane wind speeds and may be relevant to the finding that 68% of all category 5 hurricanes pass through the plume region.
Acknowledgments
The author wishes to thank the three anonymous reviewers for their insightful comments and the participants of the NBCR experiment for the success of the project.
REFERENCES
Astor, Y., F. Muller-Karger, and M. I. Scranton, 2003: Seasonal and interannual variation in the hydrography of the Cariaco Basin: Implications for basin ventilation. Cont. Shelf Res., 23 , 125–144.
Bosart, L. F., W. E. Bracken, J. Molinari, C. S. Velden, and P. G. Black, 2000: Environmental influences on the rapid intensification of Hurricane Opal (1995) over the Gulf of Mexico. Mon. Wea. Rev., 128 , 322–352.
Carpenter, E. J., A. Subramaniam, and D. G. Capone, 2004: Biomass and primary productivity of the cyanobacterium Trichodesmium spp. in the tropical N. Atlantic Ocean. Deep-Sea Res. I, 51 , 173–203.
Chang, P., R. Saravanan, L. Ji, and G. C. Hegerl, 2000: The effect of local sea surface temperatures on atmospheric circulation over the tropical Atlantic sector. J. Climate, 13 , 2195–2216.
Cione, J. J., and E. W. Uhlhorn, 2003: Sea surface temperature variability in hurricanes: Implications with respect to intensity change. Mon. Wea. Rev., 131 , 1783–1796.
Cutter, G. A., L. S. Cutter, A. M. Featherstone, and S. E. Lohrenz, 2001: Antimony and arsenic biogeochemistry in the western Atlantic Ocean. Deep-Sea Res. II, 48 , 2895–2915.
Dagg, M., R. Benner, S. Lohrenz, and D. Lawrence, 2004: Transformation of dissolved and particulate materials on continental shelves influenced by large rivers: Plume processes. Cont. Shelf Res., 24 , 833–858.
de Moura, R. L., J. L. de Figueiredo, and I. Sazima, 2001: A new parrotfish (Scaridae) from Brazil, and revalidation of Sparisoma amplum (Ranzani, 1842), Sparisoma frondosum (Agassiz, 1831), Sparisoma axillare (Steindachner, 1878), and Scarus trispinosus Valenciennes, 1840. Bull. Mar. Sci., 68 , 505–524.
Ffield, A., 2005: North Brazil Current rings viewed by TRMM Microwave Imager SST and the influence of the Amazon Plume. Deep-Sea Res. I, 52 , 137–160.
Fonseca, C. A., G. J. Goni, W. E. Johns, and E. J. D. Campos, 2004: Investigation of the North Brazil Current retroflection and North Equatorial Countercurrent variability. Geophys. Res. Lett., 31 .L21304, doi:10.1029/2004GL020054.
Fratantoni, D. M., and D. A. Glickson, 2002: North Brazil Current ring generation and evolution observed with SeaWiFS. J. Phys. Oceanogr., 32 , 1058–1074.
Fratantoni, D. M., and P. L. Richardson, 2006: The evolution and demise of North Brazil Current Rings. J. Phys. Oceanogr., 36 , 1241–1264.
Free, M., M. Bister, and K. Emanuel, 2004: Potential intensity of tropical cyclones: Comparison of results from radiosonde and reanalysis data. J. Climate, 17 , 1722–1727.
Fu, R., R. E. Dickinson, M. Chen, and H. Wang, 2001: How do tropical sea surface temperatures influence the seasonal distribution of precipitation in the equatorial Amazon? J. Climate, 14 , 4003–4026.
Fu, R., M. Chen, and H. Wang, 2004: Can continental rainfall influence climate variability over tropical Atlantic Ocean (2004–13AIRSEA). Preprints, 13th Conf. on Interactions of the Sea and Atmosphere, Portland, ME, Amer. Meteor. Soc., CD-ROM, 6.4.
Garraffo, Z. D., W. E. Johns, E. P. Chassignet, and G. J. Goni, 2003: North Brazil Current rings and transport of southern waters in a high resolution numerical simulation of the North Atlantic. Interhemispheric Water Exchange in the Atlantic Ocean, G. J. Goni and P. Malanotte-Rizzoli, Eds., Elsevier Oceanographic Series, Vol. 68, Elsevier, 375–409.
Garzoli, S. L., A. Ffield, and Q. Yao, 2003: North Brazil Current rings and the variability in the latitude of retroflection. Interhemispheric Water Exchange in the Atlantic Ocean, G. J. Goni and P. Malanotte-Rizzoli, Eds., Elsevier Oceanographic Series, Vol. 68, Elsevier, 357–373.
Goes, M., R. Molinari, I. da Silveriera, and I. Wainer, 2005: Retroflections of the North Brazil Current during February 2002. Deep-Sea Res. I, 52 , 647–667.
Goni, G. J., and W. E. Johns, 2003: Synoptic study of warm rings in the North Brazil Current retroflection region using satellite altimeter data. Interhemispheric Water Exchange in the Atlantic Ocean, G. J. Goni and P. Malanotte-Rizzoli, Eds., Elsevier Oceanographic Series, Vol. 68, Elsevier, 335–356.
Goni, G. J., and J. A. Trinanes, 2003: Ocean thermal structure monitoring could aid in the intensity forecast of tropical cyclones. Eos, Trans. Amer. Geophys. Union, 84 , 573–577, and 578.
Goni, G. J., M. DeMaria, J. A. Trinanes, and P. G. Black, 2004: Testing global estimates of the tropical cyclone heat potential fields to improve hurricane intensification prediction (2004–26HURR). Preprints, 26th Conf. on Hurricanes and Tropical Meteorology, Miami, FL, Amer. Meteor. Soc., CD-ROM, 1A.5.
Gonzalez-Silvera, A., E. Santamaria-del-Angel, V. M. T. Garcia, C. A. E. Garcia, R. Nunez-Millan, and F. Muller-Karger, 2004: Biogeographical regions of the tropical and subtropical Atlantic Ocean off South America: Classification based on pigment (CZCS) and chlorophyll-a (SeaWiFS) variability. Cont. Shelf Res., 24 , 983–1000.
Hellweger, F. L., and A. L. Gordon, 2002: Tracing Amazon River water into the Caribbean Sea. J. Mar. Res., 60 , 537–549.
Jochum, M., and P. Malanotte-Rizzoli, 2003: On the generation of North Brazil Current Rings. J. Mar. Res., 61 , 147–173.
Johns, W. E., R. J. Zantopp, and G. J. Goni, 2003: Cross-gyre transport by North Brazil Current Rings. Interhemispheric Water Exchange in the Atlantic Ocean, G. J. Goni and P. Malanotte-Rizzoli, Eds., Elsevier Oceanographic Series, Vol. 68, Elsevier, 411–441.
Kelly, P. S., K. M. M. Lwiza, R. K. Cowen, and G. J. Goni, 2000: Low-salinity pools at Barbados, West Indies: Their origin, frequency, and variability. J. Geophys. Res., 105 , 19699–19708.
Koertzinger, A., 2003: A significant CO2 sink in the tropical Atlantic Ocean associated with the Amazon River plume. Geophys. Res. Lett., 30 .2287, doi:10.1029/2003GL018841.
Labat, D., J. Ronchail, J. Callede, J. L. Guyot, E. de Oliveira, and W. Guimaraes, 2004: Wavelet analysis of Amazon hydrological regime variability. Geophys. Res. Lett., 31 .L02501, doi:10.1029/2003GL018741.
Lawrence, M. B., 2000: Tropical cyclone report, Hurricane Joyce, 25 September – 2 October 2000. National Hurricane Center, 6 pp. [Available online at http://www.nhc.noaa.gov/2000joyce.html.].
Lentz, S. J., 1995: Seasonal variations in the horizontal structure of the Amazon Plume inferred from historical hydrographic data. J. Geophys. Res., 100 , 2391–2400.
Lentz, S. J., and R. Limeburner, 1995: The Amazon River Plume during AMASSEDS: Spatial characteristics and salinity variability. J. Geophys. Res., 100 , 2355–2376.
Liebmann, B., and J. A. Marengo, 2001: Interannual variability of the rainy season and rainfall in the Brazilian Amazon basin. J. Climate, 14 , 4308–4318.
Limeburner, R., R. C. Beardsley, I. D. Soares, S. J. Lentz, and J. Candela, 1995: Lagrangian flow observations of the Amazon River discharge into the North Atlantic. J. Geophys. Res., 100 , 2401–2416.
Lin, I-I., W. T. Liu, C-C. Wu, J. C. H. Chiang, and C-H. Sui, 2003: Satellite observations of modulation of surface winds by typhoon-induced upper ocean cooling. Geophys. Res. Lett., 30 .1131, doi:10.1029/2002GL015674.
Muller-Karger, F. E., C. R. McClain, and P. L. Richardson, 1988: The dispersal of the Amazon’s water. Nature, 333 , 56–59.
Ou, H. W., 1989: Why does the Amazon water flow to the north after its discharge? J. Phys. Oceanogr., 19 , 1102–1107.
Pailler, K., B. Bourlès, and Y. Gouriou, 1999: The barrier layer in the western tropical Atlantic Ocean. Geophys. Res. Lett., 26 , 2069–2072.
Perrie, W., W. Zhang, Z. Long, and E. L. Andreas, 2004: Impacts of waves, sea spray, and the upper ocean on extratropical storms (2004 - 13AIRSEA). Preprints, 13th Conf. on the Interactions of Sea and Atmosphere, Portland, ME, Amer. Meteor. Soc., CD-ROM, 8.5.
Perry, G. D., P. B. Duffy, and N. L. Miller, 1996: An extended data set of river discharges for validation of general circulation models. J. Geophys. Res., 101 , 21339–21349.
Ronchail, J., G. Cochonneau, M. Molinier, J. L. Guyot, A. G. De Miranda Chaves, V. Guimaraes, and E. De Oliveira, 2002: Interannual rainfall variability in the Amazon Basin and sea-surface temperatures in the Equatorial Pacific and the Tropical Atlantic Oceans. Int. J. Climatol., 22 , 1663–1686.
Shay, L. K., G. J. Goni, and P. G. Black, 2000: Effects of a warm oceanic feature on Hurricane Opal. Mon. Wea. Rev., 128 , 1366–1383.
Shay, L. K., T. Cook, R. Evans, E. W. Uhlhorn, S. Guhin, J. Cione, S. White, and M. L. Black, 2004: Observed modulation of upper ocean heat content variability by upper ocean current shears during Hurricanes Isidore and Lili (2004–26HURR). Preprints, 26th Conf. on Hurricanes and Tropical Meteorology, Miami, FL, Amer. Meteor. Soc., CD-ROM, 16A.4.
Signorini, S. R., R. G. Murtugudde, C. R. McClain, J. R. Christian, J. Picaut, and A. J. Busalacchi, 1999: Biological and physical signatures in the tropical and subtropical Atlantic. J. Geophys. Res., 104 , 18367–18382.
Stewart, S. R., 2005: Tropical cyclone report, Hurricane Ivan, 2–26 September 2004. National Hurricane Center, 40 pp. [Available online at http://www.nhc.noaa.gov/2004ivan.shtml.].
Ternon, J. F., C. Oudot, A. Dessier, and D. Diverres, 2000: A seasonal tropical sink for atmospheric CO2 in the Atlantic Ocean: The role of the Amazon River discharge. Mar. Chem., 68 , 183–201.
Wang, C., and D. B. Enfield, 2001: The tropical Western Hemisphere warm pool. Geophys. Res. Lett., 28 , 1635–1638.
Wang, H., and R. Fu, 2002: Cross-equatorial flow and seasonal cycle of precipitation over South America. J. Climate, 15 , 1591–1608.
Warne, A. G., R. H. Meade, W. A. White, E. H. Guevara, J. Gibeaut, R. C. Smyth, A. Aslan, and T. Tremblay, 2002: Regional controls on geomorphology, hydrology, and ecosystem integrity in the Orinoco Delta. Geomorphology, 44 , 273–307.
Wentz, F. J., and T. Meissner, 2000: Algorithm Theoretical Basis Document (ATBD), Version 2. AMSR Ocean Algorithm, RSS Tech. Proposal 121599A-1, Remote Sensing Systems, Santa Rosa, CA, 59 pp.
Wilson, D. W., W. E. Johns, and S. L. Garzoli, 2002: Velocity structure of North Brazil Current rings. Geophys. Res. Lett., 29 .1273, doi:10.1029/2001GL013869.
The June through November hurricane season: (a) average surface salinity highlighting the Amazon–Orinoco River plume signature, (b) surface temperature for only profiles with surface salinities less than 33.5 to isolate the temperature signal of the Amazon–Orinoco River plume, (c) average surface temperature, and (d) surface temperature extremes with only temperatures greater than 29°C (primarily located in the plume region) and less than 26°C (found in the NBC and the northeastern quadrant) plotted.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The June through November hurricane season: (a) average surface salinity highlighting the Amazon–Orinoco River plume signature, (b) surface temperature for only profiles with surface salinities less than 33.5 to isolate the temperature signal of the Amazon–Orinoco River plume, (c) average surface temperature, and (d) surface temperature extremes with only temperatures greater than 29°C (primarily located in the plume region) and less than 26°C (found in the NBC and the northeastern quadrant) plotted.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The June through November hurricane season: (a) average surface salinity highlighting the Amazon–Orinoco River plume signature, (b) surface temperature for only profiles with surface salinities less than 33.5 to isolate the temperature signal of the Amazon–Orinoco River plume, (c) average surface temperature, and (d) surface temperature extremes with only temperatures greater than 29°C (primarily located in the plume region) and less than 26°C (found in the NBC and the northeastern quadrant) plotted.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The 1960 to 1965 hurricane tracks for tropical cyclones that attained maximum wind speeds of 83 kt or greater (category 2 hurricane or higher). The historical average extent of the Amazon–Orinoco River plume during hurricane season (Fig. 1a) is shown in green. Hurricane tracks passing through (outside) the plume are indicated by thick yellow (thin blue) lines. The tropical cyclones initially travel westward. The southernmost hurricane track is Hurricane Flora, 1963, a category 4 hurricane. The historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is shown in red. Schematic positions of the NBC, the NBC retroflected flow (RTFL), and a recently shed NBC ring (R) are indicated on the map.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The 1960 to 1965 hurricane tracks for tropical cyclones that attained maximum wind speeds of 83 kt or greater (category 2 hurricane or higher). The historical average extent of the Amazon–Orinoco River plume during hurricane season (Fig. 1a) is shown in green. Hurricane tracks passing through (outside) the plume are indicated by thick yellow (thin blue) lines. The tropical cyclones initially travel westward. The southernmost hurricane track is Hurricane Flora, 1963, a category 4 hurricane. The historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is shown in red. Schematic positions of the NBC, the NBC retroflected flow (RTFL), and a recently shed NBC ring (R) are indicated on the map.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The 1960 to 1965 hurricane tracks for tropical cyclones that attained maximum wind speeds of 83 kt or greater (category 2 hurricane or higher). The historical average extent of the Amazon–Orinoco River plume during hurricane season (Fig. 1a) is shown in green. Hurricane tracks passing through (outside) the plume are indicated by thick yellow (thin blue) lines. The tropical cyclones initially travel westward. The southernmost hurricane track is Hurricane Flora, 1963, a category 4 hurricane. The historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is shown in red. Schematic positions of the NBC, the NBC retroflected flow (RTFL), and a recently shed NBC ring (R) are indicated on the map.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The number of 1950 through 2003 “best track” tropical storms and hurricanes per one degree square (smoothed by a 3° × 3° block average). The tropical cyclones initially travel westward.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The number of 1950 through 2003 “best track” tropical storms and hurricanes per one degree square (smoothed by a 3° × 3° block average). The tropical cyclones initially travel westward.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The number of 1950 through 2003 “best track” tropical storms and hurricanes per one degree square (smoothed by a 3° × 3° block average). The tropical cyclones initially travel westward.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Scatterplot of the June–November SST vs sea surface salinity for a zonal strip frequently traversed by tropical cyclones. All points with salinities greater than 35.8 are colored blue, and all points with salinities less than 35.8 are colored red. The saltier salinities (blue) are found to the east (map inset) and on average they are 2°C cooler than the fresher salinities, which are located to the west, associated with the Amazon–Orinoco River plume. Tropical cyclones traversing this zonal strip may abruptly encounter considerably warmer plume SSTs as they initially move westward. The historical average extent of the Amazon–Orinoco River plume during hurricane season (Fig. 1a) is shown in green. The historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is contoured in red. Schematic positions of the NBC, the NBC retroflected flow (RTFL), and a recently shed NBC ring (R) are indicated on the map.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Scatterplot of the June–November SST vs sea surface salinity for a zonal strip frequently traversed by tropical cyclones. All points with salinities greater than 35.8 are colored blue, and all points with salinities less than 35.8 are colored red. The saltier salinities (blue) are found to the east (map inset) and on average they are 2°C cooler than the fresher salinities, which are located to the west, associated with the Amazon–Orinoco River plume. Tropical cyclones traversing this zonal strip may abruptly encounter considerably warmer plume SSTs as they initially move westward. The historical average extent of the Amazon–Orinoco River plume during hurricane season (Fig. 1a) is shown in green. The historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is contoured in red. Schematic positions of the NBC, the NBC retroflected flow (RTFL), and a recently shed NBC ring (R) are indicated on the map.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Scatterplot of the June–November SST vs sea surface salinity for a zonal strip frequently traversed by tropical cyclones. All points with salinities greater than 35.8 are colored blue, and all points with salinities less than 35.8 are colored red. The saltier salinities (blue) are found to the east (map inset) and on average they are 2°C cooler than the fresher salinities, which are located to the west, associated with the Amazon–Orinoco River plume. Tropical cyclones traversing this zonal strip may abruptly encounter considerably warmer plume SSTs as they initially move westward. The historical average extent of the Amazon–Orinoco River plume during hurricane season (Fig. 1a) is shown in green. The historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is contoured in red. Schematic positions of the NBC, the NBC retroflected flow (RTFL), and a recently shed NBC ring (R) are indicated on the map.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Average hurricane season (a) temperature and (b) salinity ocean profiles, with one standard deviation envelopes, for a zonal strip that is frequently traversed by tropical cyclones (same zonal strip as used in Fig. 4), and a region to the north (colored green). Within the zonal strip, all profiles with surface salinities greater than 35.8 are colored blue, and all profiles with surface salinities less than 35.8 are colored red (river-influenced plume water). The plume waters (red) are significantly warmer throughout the upper 100 db (∼m) and clearly fresher down to 50 m than the cooler and saltier waters to the east (blue) and to the north (green). Therefore, tropical cyclones can abruptly encounter considerably warmer upper-ocean temperatures in the plume, even if the ocean becomes well mixed down to 100 m. On the inset station map, the historical average extent of the Amazon–Orinoco River plume during hurricane season (Fig. 1a) is shown in green, and the historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is contoured in red.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Average hurricane season (a) temperature and (b) salinity ocean profiles, with one standard deviation envelopes, for a zonal strip that is frequently traversed by tropical cyclones (same zonal strip as used in Fig. 4), and a region to the north (colored green). Within the zonal strip, all profiles with surface salinities greater than 35.8 are colored blue, and all profiles with surface salinities less than 35.8 are colored red (river-influenced plume water). The plume waters (red) are significantly warmer throughout the upper 100 db (∼m) and clearly fresher down to 50 m than the cooler and saltier waters to the east (blue) and to the north (green). Therefore, tropical cyclones can abruptly encounter considerably warmer upper-ocean temperatures in the plume, even if the ocean becomes well mixed down to 100 m. On the inset station map, the historical average extent of the Amazon–Orinoco River plume during hurricane season (Fig. 1a) is shown in green, and the historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is contoured in red.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Average hurricane season (a) temperature and (b) salinity ocean profiles, with one standard deviation envelopes, for a zonal strip that is frequently traversed by tropical cyclones (same zonal strip as used in Fig. 4), and a region to the north (colored green). Within the zonal strip, all profiles with surface salinities greater than 35.8 are colored blue, and all profiles with surface salinities less than 35.8 are colored red (river-influenced plume water). The plume waters (red) are significantly warmer throughout the upper 100 db (∼m) and clearly fresher down to 50 m than the cooler and saltier waters to the east (blue) and to the north (green). Therefore, tropical cyclones can abruptly encounter considerably warmer upper-ocean temperatures in the plume, even if the ocean becomes well mixed down to 100 m. On the inset station map, the historical average extent of the Amazon–Orinoco River plume during hurricane season (Fig. 1a) is shown in green, and the historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is contoured in red.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Two contrasting examples of the anticyclonic (clockwise) NBC rings are shown with depth vs distance (km; northern end at 0 km) sections for temperature (°C), salinity, and velocity (cm s−1; westward: −; eastward: +). (a) A thermocline-intensified ring observed in December 1998 that reveals maximum swirl velocities below 100 m, and (b) a surface-intensified ring observed in February 1999 that reveals maximum swirl velocities above 75 m. In (a) warm (>28.4°C) and low-salinity (<35.4) Amazon–Orinoco River plume water is revealed in the topmost 30–40 m centered over the core of the spinning ring. In (b), high-salinity (>36.2) South Atlantic water is observed throughout the ring’s upper layers with low-salinity (<35.4) plume waters caught only on the edges of the ring in the topmost 30 m.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Two contrasting examples of the anticyclonic (clockwise) NBC rings are shown with depth vs distance (km; northern end at 0 km) sections for temperature (°C), salinity, and velocity (cm s−1; westward: −; eastward: +). (a) A thermocline-intensified ring observed in December 1998 that reveals maximum swirl velocities below 100 m, and (b) a surface-intensified ring observed in February 1999 that reveals maximum swirl velocities above 75 m. In (a) warm (>28.4°C) and low-salinity (<35.4) Amazon–Orinoco River plume water is revealed in the topmost 30–40 m centered over the core of the spinning ring. In (b), high-salinity (>36.2) South Atlantic water is observed throughout the ring’s upper layers with low-salinity (<35.4) plume waters caught only on the edges of the ring in the topmost 30 m.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Two contrasting examples of the anticyclonic (clockwise) NBC rings are shown with depth vs distance (km; northern end at 0 km) sections for temperature (°C), salinity, and velocity (cm s−1; westward: −; eastward: +). (a) A thermocline-intensified ring observed in December 1998 that reveals maximum swirl velocities below 100 m, and (b) a surface-intensified ring observed in February 1999 that reveals maximum swirl velocities above 75 m. In (a) warm (>28.4°C) and low-salinity (<35.4) Amazon–Orinoco River plume water is revealed in the topmost 30–40 m centered over the core of the spinning ring. In (b), high-salinity (>36.2) South Atlantic water is observed throughout the ring’s upper layers with low-salinity (<35.4) plume waters caught only on the edges of the ring in the topmost 30 m.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Two microwave satellite-derived SST composite images of the Amazon plume region revealing the SST conditions (a) before and (b) after the passing of Hurricane Joyce, a category 1 hurricane that attained wind speeds of 80 kt in September–October 2000. Color-coded circles mark the 6-h positions and wind speeds of the hurricane. A white outlined box centered at 10°N, 50°W highlights the NBC ring generation region and is shown for easier cross comparison. Five days of data, centered on 23 Sep 2000 in (a) and 2 Oct 2000 in (b), have been averaged to construct the SST images, which are smoothed by a 2° × 2° block average.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Two microwave satellite-derived SST composite images of the Amazon plume region revealing the SST conditions (a) before and (b) after the passing of Hurricane Joyce, a category 1 hurricane that attained wind speeds of 80 kt in September–October 2000. Color-coded circles mark the 6-h positions and wind speeds of the hurricane. A white outlined box centered at 10°N, 50°W highlights the NBC ring generation region and is shown for easier cross comparison. Five days of data, centered on 23 Sep 2000 in (a) and 2 Oct 2000 in (b), have been averaged to construct the SST images, which are smoothed by a 2° × 2° block average.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Two microwave satellite-derived SST composite images of the Amazon plume region revealing the SST conditions (a) before and (b) after the passing of Hurricane Joyce, a category 1 hurricane that attained wind speeds of 80 kt in September–October 2000. Color-coded circles mark the 6-h positions and wind speeds of the hurricane. A white outlined box centered at 10°N, 50°W highlights the NBC ring generation region and is shown for easier cross comparison. Five days of data, centered on 23 Sep 2000 in (a) and 2 Oct 2000 in (b), have been averaged to construct the SST images, which are smoothed by a 2° × 2° block average.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The westward translation speed time series of (a) relatively slow-moving Hurricane Joyce and (b) relatively fast-moving Hurricane Ivan, calculated as degrees of longitude per day; the averages are indicated by the straight solid lines. (c), (d) The SST (°C) difference maps, calculated as the difference in SST between the “SST after” and “SST before” maps of the corresponding figures (Fig. 7 and Fig. 9). (c) For Hurricane Joyce, the SST difference map reveals the sharply delineated cool NBC ring (−0.9° to −1.5°C), and (d) for Hurricane Ivan, the SST difference map reveals unstructured, yet overall cool temperatures (down to −1.0°C). In (c) and (d), the white outlined boxes centered at 10°N, 50°W highlight the NBC ring generation region and are shown for easier cross comparison.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The westward translation speed time series of (a) relatively slow-moving Hurricane Joyce and (b) relatively fast-moving Hurricane Ivan, calculated as degrees of longitude per day; the averages are indicated by the straight solid lines. (c), (d) The SST (°C) difference maps, calculated as the difference in SST between the “SST after” and “SST before” maps of the corresponding figures (Fig. 7 and Fig. 9). (c) For Hurricane Joyce, the SST difference map reveals the sharply delineated cool NBC ring (−0.9° to −1.5°C), and (d) for Hurricane Ivan, the SST difference map reveals unstructured, yet overall cool temperatures (down to −1.0°C). In (c) and (d), the white outlined boxes centered at 10°N, 50°W highlight the NBC ring generation region and are shown for easier cross comparison.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The westward translation speed time series of (a) relatively slow-moving Hurricane Joyce and (b) relatively fast-moving Hurricane Ivan, calculated as degrees of longitude per day; the averages are indicated by the straight solid lines. (c), (d) The SST (°C) difference maps, calculated as the difference in SST between the “SST after” and “SST before” maps of the corresponding figures (Fig. 7 and Fig. 9). (c) For Hurricane Joyce, the SST difference map reveals the sharply delineated cool NBC ring (−0.9° to −1.5°C), and (d) for Hurricane Ivan, the SST difference map reveals unstructured, yet overall cool temperatures (down to −1.0°C). In (c) and (d), the white outlined boxes centered at 10°N, 50°W highlight the NBC ring generation region and are shown for easier cross comparison.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Two microwave satellite-derived SST composite images of the Amazon plume region revealing the SST conditions (a) before and (b) after the passing of Hurricane Ivan, a category 5 hurricane that attained wind speeds of 140 kt in September 2004. Color-coded circles mark the 6-h positions and wind speeds of the hurricane. A white outlined box centered at 10°N, 50°W highlights the NBC ring generation region and is shown for easier cross comparison. Five days of data, centered on (a) 1 Sep 2004 and (b) 8 Sep 2004 have been averaged to construct the SST images, which are smoothed by a 2° × 2° block average.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Two microwave satellite-derived SST composite images of the Amazon plume region revealing the SST conditions (a) before and (b) after the passing of Hurricane Ivan, a category 5 hurricane that attained wind speeds of 140 kt in September 2004. Color-coded circles mark the 6-h positions and wind speeds of the hurricane. A white outlined box centered at 10°N, 50°W highlights the NBC ring generation region and is shown for easier cross comparison. Five days of data, centered on (a) 1 Sep 2004 and (b) 8 Sep 2004 have been averaged to construct the SST images, which are smoothed by a 2° × 2° block average.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Two microwave satellite-derived SST composite images of the Amazon plume region revealing the SST conditions (a) before and (b) after the passing of Hurricane Ivan, a category 5 hurricane that attained wind speeds of 140 kt in September 2004. Color-coded circles mark the 6-h positions and wind speeds of the hurricane. A white outlined box centered at 10°N, 50°W highlights the NBC ring generation region and is shown for easier cross comparison. Five days of data, centered on (a) 1 Sep 2004 and (b) 8 Sep 2004 have been averaged to construct the SST images, which are smoothed by a 2° × 2° block average.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
(a) The time series of the June–November average tropical cyclone wind speed (“+” symbols) with the time series of the June–November average infrared satellite-derived SST (“○” symbols) from 7.5°N, 52.5°W. The wind speed time series is averaged from all the tropical cyclone track data in the 7.9°–25.0°N, 55.0°–35.0°W block. The SST time series is from the location at the northern edge of the Guiana Plateau where the highest cross-correlation coefficient, 0.67, is observed in Fig. 11. (b) The 1-yr block filtered Niño-3 (5°N–5°S, 150°–90°W) SST index plotted to show ENSO variability for comparison.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
(a) The time series of the June–November average tropical cyclone wind speed (“+” symbols) with the time series of the June–November average infrared satellite-derived SST (“○” symbols) from 7.5°N, 52.5°W. The wind speed time series is averaged from all the tropical cyclone track data in the 7.9°–25.0°N, 55.0°–35.0°W block. The SST time series is from the location at the northern edge of the Guiana Plateau where the highest cross-correlation coefficient, 0.67, is observed in Fig. 11. (b) The 1-yr block filtered Niño-3 (5°N–5°S, 150°–90°W) SST index plotted to show ENSO variability for comparison.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
(a) The time series of the June–November average tropical cyclone wind speed (“+” symbols) with the time series of the June–November average infrared satellite-derived SST (“○” symbols) from 7.5°N, 52.5°W. The wind speed time series is averaged from all the tropical cyclone track data in the 7.9°–25.0°N, 55.0°–35.0°W block. The SST time series is from the location at the northern edge of the Guiana Plateau where the highest cross-correlation coefficient, 0.67, is observed in Fig. 11. (b) The 1-yr block filtered Niño-3 (5°N–5°S, 150°–90°W) SST index plotted to show ENSO variability for comparison.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Map of the cross-correlation coefficient between the 1982–2003 June–November averaged (one value per year) tropical cyclone wind speed time series for the region outlined by the black rectangle and each 1982–2003 June–November averaged (one value per year) infrared satellite-derived SST time series for each one-degree square. The solid black arrow is the average position of the tropical cyclone track positions within the region used to create the wind speed time series (7.9°–25.0°N, 55.0°–35.0°W; southernmost tropical cyclone track is at 7.9°N; the time series is shown in Fig. 10a; “+” symbols). The historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is contoured in red. The higher cross-correlation coefficients, >0.5, are almost all located within the historical position of the plume. The highest cross-correlation coefficient, 0.67, is located at 7.5°N, 52.5°W (marked with a white star), in the region where recently shed NBC rings typically pass through.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Map of the cross-correlation coefficient between the 1982–2003 June–November averaged (one value per year) tropical cyclone wind speed time series for the region outlined by the black rectangle and each 1982–2003 June–November averaged (one value per year) infrared satellite-derived SST time series for each one-degree square. The solid black arrow is the average position of the tropical cyclone track positions within the region used to create the wind speed time series (7.9°–25.0°N, 55.0°–35.0°W; southernmost tropical cyclone track is at 7.9°N; the time series is shown in Fig. 10a; “+” symbols). The historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is contoured in red. The higher cross-correlation coefficients, >0.5, are almost all located within the historical position of the plume. The highest cross-correlation coefficient, 0.67, is located at 7.5°N, 52.5°W (marked with a white star), in the region where recently shed NBC rings typically pass through.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
Map of the cross-correlation coefficient between the 1982–2003 June–November averaged (one value per year) tropical cyclone wind speed time series for the region outlined by the black rectangle and each 1982–2003 June–November averaged (one value per year) infrared satellite-derived SST time series for each one-degree square. The solid black arrow is the average position of the tropical cyclone track positions within the region used to create the wind speed time series (7.9°–25.0°N, 55.0°–35.0°W; southernmost tropical cyclone track is at 7.9°N; the time series is shown in Fig. 10a; “+” symbols). The historical average extent of the warmest temperatures of the Amazon–Orinoco River plume during hurricane season (Fig. 1b) is contoured in red. The higher cross-correlation coefficients, >0.5, are almost all located within the historical position of the plume. The highest cross-correlation coefficient, 0.67, is located at 7.5°N, 52.5°W (marked with a white star), in the region where recently shed NBC rings typically pass through.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The correlation coefficients for the linear regressions (not shown) of all the lagged sets of 52 pairs of values of the average wind speed for each individual tropical storm and hurricane passing through the 7.9°–25.0°N, 55.0°–35.0°W block for the years 1982 through 2003, and the infrared satellite-derived SST from 7.5°N, 52.5°W identified at the same time as each event, and as a function of lagged time, for each event. The largest correlation coefficients are associated with SSTs 14.6 days earlier than wind speeds, possibly associated with the fortnightly tides and with SSTs 51.1–54.8 days earlier than the wind speeds, and possibly associated with NBC rings.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The correlation coefficients for the linear regressions (not shown) of all the lagged sets of 52 pairs of values of the average wind speed for each individual tropical storm and hurricane passing through the 7.9°–25.0°N, 55.0°–35.0°W block for the years 1982 through 2003, and the infrared satellite-derived SST from 7.5°N, 52.5°W identified at the same time as each event, and as a function of lagged time, for each event. The largest correlation coefficients are associated with SSTs 14.6 days earlier than wind speeds, possibly associated with the fortnightly tides and with SSTs 51.1–54.8 days earlier than the wind speeds, and possibly associated with NBC rings.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The correlation coefficients for the linear regressions (not shown) of all the lagged sets of 52 pairs of values of the average wind speed for each individual tropical storm and hurricane passing through the 7.9°–25.0°N, 55.0°–35.0°W block for the years 1982 through 2003, and the infrared satellite-derived SST from 7.5°N, 52.5°W identified at the same time as each event, and as a function of lagged time, for each event. The largest correlation coefficients are associated with SSTs 14.6 days earlier than wind speeds, possibly associated with the fortnightly tides and with SSTs 51.1–54.8 days earlier than the wind speeds, and possibly associated with NBC rings.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The four-year running Gaussian mean is removed from the 7.5°N, 52.5°W infrared satellite-derived SST time series and the values from 53 days earlier are paired with average tropical cyclone track wind speed anomaly values from the 7.9°–25.0°N, 55.0°–35.0°W region. The scatterplot of the anomaly pairs reveals that earlier, warm SSTs at the northern edge of the Guiana Plateau are associated with both weak and strong tropical cyclone wind speeds, but earlier, cool SSTs—possibly from recently shed NBC rings—are almost always (never) associated with weak (strong) average tropical cyclone wind speeds.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The four-year running Gaussian mean is removed from the 7.5°N, 52.5°W infrared satellite-derived SST time series and the values from 53 days earlier are paired with average tropical cyclone track wind speed anomaly values from the 7.9°–25.0°N, 55.0°–35.0°W region. The scatterplot of the anomaly pairs reveals that earlier, warm SSTs at the northern edge of the Guiana Plateau are associated with both weak and strong tropical cyclone wind speeds, but earlier, cool SSTs—possibly from recently shed NBC rings—are almost always (never) associated with weak (strong) average tropical cyclone wind speeds.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The four-year running Gaussian mean is removed from the 7.5°N, 52.5°W infrared satellite-derived SST time series and the values from 53 days earlier are paired with average tropical cyclone track wind speed anomaly values from the 7.9°–25.0°N, 55.0°–35.0°W region. The scatterplot of the anomaly pairs reveals that earlier, warm SSTs at the northern edge of the Guiana Plateau are associated with both weak and strong tropical cyclone wind speeds, but earlier, cool SSTs—possibly from recently shed NBC rings—are almost always (never) associated with weak (strong) average tropical cyclone wind speeds.
Citation: Journal of Climate 20, 2; 10.1175/JCLI3985.1
The distribution of 1960–2000 hurricanes by location. With increasing category (hurricane strength), an increasing (decreasing) percentage of hurricanes pass through (outside) the plume region. For example, for category 5 hurricanes, 68% passed through the plume region, while only 32% passed outside the plume region.
The distribution of 1960–2000 hurricanes by category within each location. With increasing category (hurricane strength), the percentage of hurricanes passing through (outside) the plume region remains relatively similar (decreases). For example, for hurricanes passing through the plume region, 16% are category 5 and relatively similarly, 22% are category 1, whereas for hurricanes passing outside the plume region, only 4% are category 5 and significantly more, 54%, are category 1.
