Feedbacks between Hydrological Processes in Tropical South America and Large-Scale Ocean–Atmospheric Phenomena

Germán Poveda Postgrado en Aprovechamiento de Recursos Hidráulicos, Facultad de Minas, Universidad Nacional de Colombia, Medellin, Colombia

Search for other papers by Germán Poveda in
Current site
Google Scholar
PubMed
Close
and
Oscar J. Mesa Postgrado en Aprovechamiento de Recursos Hidráulicos, Facultad de Minas, Universidad Nacional de Colombia, Medellin, Colombia

Search for other papers by Oscar J. Mesa in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

The hydroclimatology of tropical South America is strongly coupled to low-frequency large-scale oceanicand atmospheric phenomena occurring over the Pacific and the Atlantic Oceans. In particular, El Niño–SouthernOscillation (ENSO) affects climatic and hydrologic conditions on timescales ranging from seasons to decades.With some regional differences in timing and amplitude, tropical South America exhibits negative rainfall andstreamflow anomalies in association with the low–warm phase of the Southern Oscillation (El Niño), and positiveanomalies with the high–cold phase. Such dependence is illustrated in the hydroclimatology of Colombia throughseveral empirical analyses: correlation, empirical orthogonal functions, principal component, and spectral analysis, and discussion of the major physical mechanisms. Observations show that ENSO’s effect on river dischargesoccurs progressively later for rivers toward the east in Colombia and northern South America. Also, the impactsof La Niña are more pronounced than those of El Niño. Evidence is also presented to show that processes arisingfrom land–atmosphere interactions in tropical South America affect sea surface temperatures in the Caribbeanand the north tropical Atlantic. A hypothesis is formulated to explain these feedback mechanisms throughperturbations in precipitation, soil moisture, and evapotranspiration over the continent. To begin with, the occurrence of both phases of ENSO affects all those fields. The proposed mechanisms would constitute the “land–atmosphere” bridge connecting Pacific and Atlantic SST anomalies.

Corresponding author address: Dr. Germán Poveda, Facultad deMinas, Universidad Nacional de Colombia, 1027 A. A. Medellín,Colombia.

Email: gpoveda@perseus.unalmed.edu.co

Abstract

The hydroclimatology of tropical South America is strongly coupled to low-frequency large-scale oceanicand atmospheric phenomena occurring over the Pacific and the Atlantic Oceans. In particular, El Niño–SouthernOscillation (ENSO) affects climatic and hydrologic conditions on timescales ranging from seasons to decades.With some regional differences in timing and amplitude, tropical South America exhibits negative rainfall andstreamflow anomalies in association with the low–warm phase of the Southern Oscillation (El Niño), and positiveanomalies with the high–cold phase. Such dependence is illustrated in the hydroclimatology of Colombia throughseveral empirical analyses: correlation, empirical orthogonal functions, principal component, and spectral analysis, and discussion of the major physical mechanisms. Observations show that ENSO’s effect on river dischargesoccurs progressively later for rivers toward the east in Colombia and northern South America. Also, the impactsof La Niña are more pronounced than those of El Niño. Evidence is also presented to show that processes arisingfrom land–atmosphere interactions in tropical South America affect sea surface temperatures in the Caribbeanand the north tropical Atlantic. A hypothesis is formulated to explain these feedback mechanisms throughperturbations in precipitation, soil moisture, and evapotranspiration over the continent. To begin with, the occurrence of both phases of ENSO affects all those fields. The proposed mechanisms would constitute the “land–atmosphere” bridge connecting Pacific and Atlantic SST anomalies.

Corresponding author address: Dr. Germán Poveda, Facultad deMinas, Universidad Nacional de Colombia, 1027 A. A. Medellín,Colombia.

Email: gpoveda@perseus.unalmed.edu.co

1. Introduction

The annual distribution of rainfall over tropical SouthAmerica is primarily influenced by the position of theintertropical convergence zone (ITCZ). The main controls of the rain space distribution are the presence ofthe Andes mountains and the eastern Pacific, and western Atlantic Oceans, the atmospheric circulation overthe Amazon basin, and vegetation and soil moisturecontrasts. Large quantities of precipitation, evapotranspiration, soil moisture, and runoff are present in tropicalSouth America, as compared with world averages. Theregion is a major center of convective activity, mostlydeveloped within large cumulonimbus clouds, fromwhich latent heat is continuously released into the atmosphere thus influencing the Hadley cells and overallglobal circulation (Riehl and Malkus 1958). The excessof precipitation over evapotranspiration in the region issuch that the combined runoffs of the Amazon, Orinoco,and Magdalena rivers account for 18.3% of the totalinflow to the world oceans (Baumgartner and Reichel1975, 95). The Atrato River in Colombia drains 35702km2 of the wettest areas of the planet, producing a meanannual discharge of 4557 m3 s−1 and an equivalent runoff of 127.6 L s−1 km−2, 5–6 times larger than the Amazon. The origin of this highly wet region over northwestern South America lies in the low-level westerlyflow from the Pacific Ocean over inland Colombia.These winds are colder and moister than the dominanteasterly trades from the Atlantic and the Caribbean(López and Howell 1967). The confluence of the twowinds, combined with the effects of surface warmingand orographic lifting, produces a highly unstable atmospheric profile causing strong convection and heavyprecipitation along the Pacific coast and western flanksof the Cordillera Occidental. This region is favored fordevelopment of tropical mesoscale convective complexes (Velasco and Fritsch 1987).

On interannual timescales, low-frequency large-scaleocean–atmosphere phenomena are highly coupled to thehydroclimatology of the region. In particular, El Niño–Southern Oscillation (ENSO) is a major forcing mechanism of climatic and hydrological anomalies. Thephysics of ENSO and its climatic consequences can befound in Horel and Wallace (1981), van Loon and Madden (1981), Rasmusson and Carpenter (1982), Glantzet al. (1991), Ropelewski and Halpert (1987), Rasmusson (1991), Trenberth (1991), and Diaz and Markgraff(1992).

The purpose of this paper is twofold: 1) to study theinterannual variability of hydrologic anomalies of rainfall and river discharges in tropical South America, particularly Colombia and the Amazon River basin, and toidentify their relationship with ENSO, and 2) to presentevidence suggesting that hydrological processes in theregion can influence oceanic–atmospheric interactionsin the Caribbean and northern tropical Atlantic. We hypothesize a model to explain the nature of this feedbackwith the support of independent observations. In conclusion, the implications of long-term oscillations, suchas ENSO, for regional hydroclimatology and water resources planning, are discussed.

2. Empirical analyses

Overall, there is a coherent pattern of hydrologicalanomalies in tropical South America during extremephases of ENSO. This is clear in regions of Costa Rica(Waylen et al. 1996); Panama (Estoque et al. 1985);Colombia (Poveda and Mesa 1993; Poveda 1994; Poveda and Mesa 1995), Venezuela (Pulwarty et al. 1992),some regions of Ecuador (Gessler 1995), Amazon basinrainfalls and discharges (Vörösmarty et al. 1996; Kousky and Kayano 1994; Marengo and Hastenrath 1993;Marengo 1992; Obregón and Nobre 1990; Richey et al.1989; Lau and Sheu 1988), and other regions of Brazil(Kousky et al. 1984; Rao and Hada 1990; Hastenrathand Greischar 1993; Kayano et al. 1988; Chu 1991).Aceituno (1988, 1989), Rogers (1988), Kiladis and Diaz(1989), Hastenrath (1976, 1990), and Halpert and Ropelewski (1992) place this region in a broader hemispherical and global context. Generally, negative anomalies in rainfall and streamflows are associated with thewarm phase of ENSO (El Niño), and positive anomalieswith the cold phase (La Niña), although there are someregional differences in timing and amplitude. As an illustration, Fig. 1 shows smoothed monthly values of theSouthern Oscillation index (SOI, defined as the standardized difference between Tahiti and Darwin sea levelpressures), the discharges of the Cauca River, Colombia(1946–94), and the average rainfall over Amazon basin(5.965 × 106 km2). The latter data are estimated fromthe set produced by the Earth Observing System Amazon Project (Instituto Nacional de Pesquisas Espaciais,Brazil, and the University of Washington, and providedby the Global Hydrology and Climate Center of theNational Aeronautics and Space Administration), whichcontains monthly rainfall gridded (0.2° lat × 0.2° long)during the period 1972–92. A 12-month low-pass digitalfilter is applied to remove higher frequencies associatedwith annual and intraannual variability. The derived correlation coefficient between SOI and the Amazon rainfall is 0.63, statistically significant at the 99.99% levelonce the degrees of freedom are reduced according tothe scale of fluctuation of the processes (Vanmarcke1988; Mesa and Poveda 1993).

To illustrate the effect of extreme phases of ENSOon Colombian hydrology, Fig. 2 (top) presents themonthly discharges of the Magdalena river at PuertoBerrío, during El Niño (triangles) and La Niña (squares)years, according to the classification given by Kiladisand Diaz (1989). The hydrologic year is considered fromJune (year 0) through May (year +1). Figure 2 (bottom)illustrates the ratio of monthly mean discharges for thewhole record and those in extreme phases of ENSO.

a. Correlation analysis

Correlation analyses are calculated between seven climatic variables over the Pacific and rainfall in Colombia. The variables used are the following: Southern Oscillation index (SOI), monthly mean sea surface temperature (SST) at regions Niño-1–2 (0°–10°S, 90°–80°W), Niño-3 (5°N–5°S, 150°–90°W), and Niño-4(5°N–5°S, 160°E–150°W), and monthly mean zonalwind velocity in those regions (Uwin). All are extractedfrom the Comprehensive Ocean–Atmosphere Data Set(COADS; Slutz et al. 1985). The series are firstsmoothed by using 12-month running means to facilitatethe examination of correlations at interannual timescales. To ease interpretation of the results (Table 1) thesigns of the SSTs and some zonal wind indices are inverted. All coefficients are significant at the 95% leveland show that Colombian precipitation is most highlycorrelated with the SOI, wind velocity at Niño-4, andSST at Niño-4 (N4), Niño-3 (N3), and Niño-1–2 (N1).Interestingly, the Niño-4 region in the central Pacificreturns higher correlations than other regions closer toSouth America. Results (not shown here) suggest thatadditional variables such as the specific humidity in theNiño-1–2 region, and the meridional gradient of SSTbetween the Colombian and Peruvian coasts are alsocorrelated to precipitation variability over western andcentral Colombia.

Cross-correlation analyses between the SOI and discharges suggest that the influence of ENSO appear earlier in the west coast of Colombia and later in the easternparts of the country. ENSO anomalies lead the hydrological anomalies by one month in the west, by 2–4months in central Colombia, and by as much as 6 monthsin eastern South America, as reported for the AmazonRiver by Richey et al. (1989) and by Eagleson (1994)for the Trombetas River. The SOI exhibits the largestcorrelation (Fig. 3) with La Vieja River (Cartago, Valledel Cauca, 4°46′N, 75°54′W) at 1-month lag (0.759),with the Luisa River (Pavo Real, Tolima, 4°13′N,75°12′W), at 2-months lag (0.59), with Sumapaz River(El Profundo, Cundinamarca 4°00′N, 74°30′W) at4-months lag (0.637), and with Chivor River (Ubalá,Boyacá, 4°47′N, 73°09′W) at 6-months lag (0.50). Allcorrelations are statistically significant at 95%. Thesestreamflow stations were chosen to have approximatelythe same latitude, but the progressive delay of the effect,as west longitude decreases, is observed everywhere.As will be discussed later, hydrological processes playa role in determining the velocity at which the anomalypropagates over the continent. Indeed, precipitation, soilmoisture, and evapotranspiration anomalies at interannual timescales interact to modulate river dischargeanomalies. Figure 4 presents a time–longitude diagramof standardized precipitation anomalies at the interannual timescale in the entire Amazon basin, for the 1972–92 period. Time series have been smoothed using a low-pass 12-month digital filter. Precipitation anomalies appear migrating eastward with time, in the period 1978through 1981. The El Niño effects during 1976–77,1982–83, 1986–87, and 1991–92 are evident in Fig. 4,as is the La Niña of 1988–89.

b. Empirical orthogonal functions analyses

EOF analysis is applied to a dataset of rainfall recordsat 88 stations distributed over Colombia (Fig. 5), whichare selected for the quality of their records between 1958and 1990. The analyses were repeated on monthly,3-month, and 12-month running averages, following removal of the annual cycle by standardization (differencefrom the monthly mean divided by the standard deviation of the corresponding month). Up to 46% of thevariance can be explained with the first four EOFs, distributed as follows:
i1520-0442-10-10-2690-eq1

Figure 6 shows the first two EOFs for the 3-monthrunning mean series of precipitation over Colombia. Thefirst EOF (Fig. 6a), explaining more than 30% of rainfallvariability, has the same sign over the entire region,suggesting coherence in regional hydrological behaviorat interannual timescales. The corresponding spatialvariability broadly represents rainfall response to ENSOforcing, with central and western regions more affectedthan those on the Caribbean coast and the east. Physicalmechanisms associated to the second EOF (Fig. 6b) arenot clear, although they could be features of distinctiverainfall forcing either from the Pacific, the Atlantic, andthe Amazon basin. The corresponding time series of thefirst principal component (PC) of rainfall (Fig. 7a) confirms that El Niño events are associated with negativerainfall anomalies and La Niña with positive ones. AFourier analysis of the first PC also identified importantspectral peaks in the frequencies associated to ENSO at54, 43, and 26 months (Fig. 7b). The origin of the spectral peak at 26 months (Fig. 7b) is an open question,although it could be a manifestation of the influence ofthe quasi-biennial oscillation (QBO). ENSO exhibits animportant quasi-biennial (QB) component, not clearlyrelated to the stratospheric QBO.

Principal components are also computed for monthlyriver discharges from 55 stations (locations in Fig. 5)in the period 1959–90. All discharges are rescaled bythe corresponding monthly standard deviations to facilitate comparison across basin scales. Basins whoserainfall inputs are affected by ENSO, show even highercorrelations to SOI than rainfall (see Fig. 1). Basins actas filters to smooth out the variability and intermittencyinherent in rainfall, but soil moisture and evapotranspiration are also affected by ENSO via temperature,wind, and moisture anomalies that vary coherently. Thelatter concept will be explored more fully.

Figure 8 maps isocorrelations between the SSTs ofthe Pacific and Indian Oceans and the first principalcomponent of Colombian river discharges. Regions ofgreatest negative correlations are the central and easternequatorial Pacific and the area of the Indian monsoon.

c. Discussion

These analyses are consistent with observations ofthe physical phenomena governing the hydroclimatology of Colombia during the extreme phases of ENSO.El Niño forcing brings low rainfall and discharges, andLa Niña is even more strongly related to higher precipitation and streamflows. Correlation analyses suggestthat ENSO may explain up to 50% of the observedvariance in Colombian hydrology beyond the annualcycle, assuming a linear relationship. Several remarksmust be made about these statistics. First, the annualcycle has been removed from both signals, by traditionalstandardization. Second, smoothed records (3–12-month low-pass filters or running means) are employedto filter out the annual cycle and higher frequencies.Third, all the analysis thus far belongs solely to thelinear domain, yet the relationship of ENSO to hydrology is highly nonlinear, and therefore aspects of thistype of dependence may be absent. The context of thissearch for cause-and-effect relationships is more complex than in a simple mechanical system, due to nonlinear interactions between different subsystems (atmosphere, ocean, soil, biosphere, cryosphere) all ofwhich change at widely diverging timescales. Finally,the statistical significance of cross correlations is clearlyaffected by the autocorrelation of each series (Brownand Katz 1991), for that reason significance figures wereadjusted according to the scale of fluctuation of the processes (Vanmarcke 1988; Mesa and Poveda 1993).

Correlations between climatic variables over the Pacific Ocean and the Colombian rainfall and runoff seriesare larger when those lead the Colombian hydrology by3–4 months. This fact provides good possibilities ofdeveloping adequate hydrological predictive models.Certainly, use of such techniques contributes to improvehydrological forecasting, with tremendous practicalconsequences.

The overall relation between the indices of ENSOand the hydrologic variables is very consistent and physically reasonable, although these statistics suggest onlyassociation and are not proof of dependence. There areinstances in which rainfall anomalies are not associatedwith extreme phases of ENSO, and vice versa. The1982–83 El Niño event is the strongest in the record,but it did not produce intense dry anomalies. During1957–60 Colombia experienced one of the more prolonged dry seasons in the record, but the 1957–58 ElNiño was not particularly marked in either duration orintensity. One of the rainiest years on record (see Fig.7a) was 1971, which was accompanied by only a moderate La Niña event. Clearly, other factors affect Colombian hydrology besides ENSO, and their dependenceis probably nonlinear. A good candidate is the NorthAtlantic oscillation (NAO), which exhibits significantcorrelations with the Colombian hydroclimatology (Poveda and Mesa 1996). Moreover, other large-scale climatic phenomena may change the degree, and even thesign of the relation. Discussion of the physical mechanisms involved follows.

3. Feedbacks

It has been observed that the north tropical Atlanticand the Caribbean experience positive, though weaker,anomalies in sea surface temperature during, or after,the warm phase of ENSO (e.g., Covey and Hastenrath1978; Hastenrath and Wu 1982; Pan and Oort 1983;Halpert and Ropelewski 1992; Curtis and Hastenrath1995; Nobre and Shukla 1996; Wagner 1996). Thiswarming tends to be stronger during the March–Mayperiod. Physical mechanisms that are responsible for thelinks between the two oceanic basins remain elusive(Lau and Nath 1994), although Curtis and Hastenrath(1995) show that the warming of the tropical Atlanticis the result of wind field perturbations.

We speculate that land–atmosphere interactions overtropical South America contribute to transmitting thesignal from the Pacific to the Caribbean and the Atlantic,once El Niño is established. Figure 9 maps correlationsbetween the first principal component of the monthlystreamflows (3-month running means, 1959–90) and theSSTs at the Caribbean and the Atlantic, for the hydrology leading the SSTs by 1–6 months. Correlations increase in both absolute value and areal extent when thehydrologic variable leads the SSTs, reaching a maximum at 4–5 months. The region of higher correlationappears to travel eastward with time. Similar correlations (not shown) for the tropical South Atlantic exhibitlower values. By comparison, lagged correlations between the SOI and the SSTs over the Atlantic (Fig. 10)are lower and less extensive. Correlations maps betweenthe Niño-3 SST record and those for the Caribbean andthe tropical North Atlantic (not shown) exhibit evenlower correlations. These results could be suggestingthat precipitation, soil moisture, and evapotranspirationare important mechanisms in establishing the importance of the South American “land–atmospherebridge,” which would connect the anomalies in the Pacific with those over the Atlantic. There is an imperfectunderstanding of the physical processes involved in suchforcing, but the overall hydrological situation of tropicalSouth America intimates strongly the existence of thebridge.

We propose the mechanisms to explain the possibleinfluence of tropical South America land–surface process on the north tropical Atlantic SST, at interannualtimescales, as follows. The anomalous warming of thePacific in the Niño-4 and Niño-3 regions acts as a Rossby wave source. The resultant changes in atmosphericcirculation produce anomalies in the global circulationthat propagate to the east as shown by Yasunari (1987see his Fig. 7 with the sea level pressure compositeanomaly) and Hsu (1994, see his Figs. 1a and 10a showing the global anomalies in OLR). Those oceanic–atmospheric perturbations attain the familiar echelon pattern throughout the central-eastern tropical Pacific, almost symmetrical about the equator, leading toward theAmericas and similar to the SST anomalies during ElNiño. Notice that the anomalies over tropical SouthAmerica exhibit an inverse sign to those of the tropicalPacific. Tropical rainfall response is quite different overthe Pacific where convection and precipitation are intensified during El Niño events, while negative anomalies prevail over continental South America.

Anomalous high surface pressure is established intropical South America [Fig. 11.27a of Gill (1982), Fig.7 of Yasunari (1987), and Fig. 5 of Aceituno (1988)],particularly during the DJF period, seemingly the resultof an anomalous Hadley cell that subdues the ascent ofmoist air and the associated convection and precipitation. This pattern appears as a common feature ofENSO, because it has been identified during El Niñoevents of 1982–83, 1986–87, and 1991–92 (Rasmussonand Mo 1993). The areal extent of those anomalies ismuch more broader than tropical South America. Thatvery observation confirms the importance of the interaction between the land and atmosphere in tropicalSouth America to influence the tropical Atlantic circulation. The impact of interannual variability of theland–atmosphere system in tropical South America isso large that the whole upper-atmospheric circulationand the divergent flux are perturbed beyond its frontiers.There are modeling results (Zeng et al. 1996) that suggest that deforestation scenarios of the Amazon basin(in many ways similar to the effects of the warm phaseof ENSO), would produce a weakening of the sea surface temperature gradient along the tropical AtlanticOcean, and also it could affect the global atmosphericcirculation through perturbations of both the Walker andthe Hadley cells (Zhang et al. 1996). Indeed, rainfall inthe Amazon basin has been recognized as a modulatorof convection in the Atlantic ITCZ and over the easternPacific (Silva Dias et al. 1987).

The anomalous increase in sea level pressure overtropical South America during El Niño contributes tothe displacement and maintenance of the center of convection of the ITCZ toward the west and the south ofits normal position (Pulwarty and Diaz 1993). DuringEl Niño, the meridional gradient of SST between Colombian coastal waters and the cold tongue diminishes.This produces a decrease in the low-level cross-equatorial westerlies winds and the advection of moisturefrom the Pacific thus contributing to the drought. Ontop of that, there is a reduction in Atlantic weather systems activity over northern South America (Frank andHebert 1974; Gray and Sheaffer 1991), thereby decreasing moisture advection and precipitation events overnorthern South America, including Colombia. Thus, ElNiño could contribute to cooling of the troposphere overthis region due to a decrease in the release of latent heatof condensation associated with negative anomalousprecipitation, despite the surface warming due to decreased evaporation and cloudiness (see Fig. 2a of Nigam 1994).

The interannual anomalies in precipitation (Lau andSheu 1988; Hsu 1994; Kousky and Kayano 1994) forcenegative anomalies in soil moisture, at the same timescale. In tropical South America this fact has been reported by Nepstad et al. (1994) and by Jipp et al. (1997),from experimental data gathered in the Amazon basinforests and pastures: soil moisture deficits associated tothe 1991–92 El Niño event persisted until 1994. Theobvious hydrological connection between soil moistureand river discharges validates the conclusions drawnfrom the isocorrelation maps shown in Fig. 9. Modelingresults confirm the interannual modulation of soil moisture at ENSO timescales. Temporal variability of monthly soil moisture anomalies in tropical South Americahas been examined through estimates of average groundwetness (a ratio of soil moisture to the maximum moisture, w/ws, ws = 150 mm) for the globe (see details inSchemm et al. 1992). Figure 11 presents the temporalevolution of monthly average ground wetness anomaliesover the Amazon, Orinoco, and Magdalena River basinsfor the period 1979–92. Data have been smoothed usinga 12-month low-pass filter (thick line). Long-term persistent behavior of soil moisture at the interannual scaleis clearly observed, as are the modulation of soil moisture by the El Niños of 1982–83 and 1986–87 and bythe La Niña of 1988.

Due to soil moisture depletion during El Niño, evapotranspiration rates are also reduced. This affects the partitioning of the surface energy balance, especially between latent and sensible heating. Nepstad et al. (1994)report a 28% reduction in evapotranspiration during anEl Niño (1992–93), compared to non–El Niño years(1991 and 1994) over Amazonian forests and a 42%decline above pastures. Evapotranspiration is clearly reduced in pastures and secondary forests of the Amazonbasin during dry seasons, in particular during thoseforced by El Niño (Hodnett et al. 1996). Similar reductions in evapotranspiration are reported by Vörösmarty et al. (1996) from water balance estimates for thewhole Amazon River basin during the period September1983–August 1984, following the 1982–1983 El Niñoevent. Interannual anomalies in soil moisture and evapotranspiration are more critical than those caused by theannual cycle. Assuming that during normal years wateris not the limiting factor to determine the Bowen ratio,the Amazon basin could be seen as an ocean. DuringENSO events, water can be the limiting factor, due tothe amplitude and length of the dry period, and thereforeits anomaly is much more critical. As a consequence,the system responds in a completely different manner.

Reductions in evapotranspiration lead to further precipitation deficits, as large proportions (35%–50%) ofrainfall in the Amazon basin are believed to be derivedfrom evapotranspiration recycling (Shuttleworth 1988;Elthair and Bras 1994). This is a crucial aspect of thefeedback mechanisms of land–atmosphere interactionsduring ENSO over tropical South America. The latentheat released into the atmosphere over the Amazon basinduring La Niña events is approximately 190 W m−2(1988–89) and reduces to 150 W m−2 in El Niño events(1991–92). Precipitation reduction is also consistentwith the development of an anomalous position and direction of the Hadley cell toward the equator. Implicithere is the existence of a positive feedback effect between the tropical precipitation and the Hadley circulation (Numaguti 1993).

During the warm phase of ENSO, the cooperativeeffect of a lower sea level pressure in the North Atlantic(Nobre and Shukla 1996) and the increased surface atmospheric pressure in tropical South America contributes to reduce the surface pressure gradient between thetwo regions. Figure 12 presents the time series of standardized anomalies for both the mean sea level pressuregradient between the Azores Islands (25°–35°W, 30°–40°N) and northern South America (50°–70°W, 3°S–10°N), as well as precipitation over the latter region, forthe period 1982–94, as obtained from the The NationalCenters for Environmental Prediction (NCEP)–NationalCenter for Atmospheric Research (NCAR) reanalysisproject. It can be seen that negative precipitation anomalies over the region are well associated to anomalousnegative sea level pressure gradient (lower North Atlantic High and higher tropical South America). Thesemechanisms in turn contribute to the weakening of thenortheast trades, triggering oceanic warming over theCaribbean and the north tropical Atlantic.

Coupling between the reduction (enhancement) of theeasterly trade winds and warming (cooling) of the tropical north Atlantic and pressure gradient reduction (increasing) during or after the warm (cold) phase of ENSOhas been reported by Horel et al. (1986), Marengo(1992), Pulwarty (1994), Curtis and Hastenrath (1995),and Nobre and Shukla (1996). Interestingly, Carton andHuang (1994) argue that during the Atlantic warmevents the ocean acts as a passive guide through whichheat is shifted from west to east in response to changesin the wind field. This kind of eastward displacementwould be manifest in the correlation maps of Fig. 9.The observations and the proposed mechanisms presented here are also in agreement with the suggestionof Zebiak (1993) that SST variability over the tropicalAtlantic may be related to land–surface interactions andto large-scale forcing related to ENSO, and also withthe proposed positive feedback between land-surfaceand atmospheric processes in the South America–Atlantic ocean–atmospheric circulation (Zeng et al. 1996).Once the SSTs increase over the Caribbean and the northtropical Atlantic, coupled ocean–atmosphere phenomena carry the signal through the rest of the Atlantictoward Africa, as suggested by Carton and Huang(1994).

4. Conclusions

Large-scale coupled ocean–atmosphere phenomenaforce hydrologic anomalies in tropical South Americaat interannual scales. Extensive data analyses illustratethe influence of ENSO on the hydrology of tropicalSouth America. Broadly, El Niño stimulates dry periodsand La Niña is associated with excessive moisture, rainfall, and river discharges. The impact of La Niña isstronger than El Niño’s impact. Because of multiplenonlinear interactions among diverse geophysical phenomena, the relationship is not a simple one, exhibitingdifferences in amplitude, timing, and duration, but thezero-order effect is clear. There is evidence of an eastward propagation of ENSO effects in streamflow in Colombian and other tropical South American rivers. Thesituation over the Atlantic is also important in modulating the influence of ENSO over the region, as theNorth Atlantic oscillation exhibits an interesting coupling with the hydrometeorology of tropical SouthAmerica. The influence of large-scale ocean–atmosphere phenomena on hydrologic anomalies needs to beaccounted for in water resources systems planning, andmanagement. In the past, during droughts associatedwith El Niño, Colombia has attempted cloud seeding tosustain reservoir levels (López 1966). The 1991–92 ElNiño event caused losses of about $1 billion to the national economy, due to prolonged electricity shortagesforced by the drought. Colombia now carefully monitorsthe evolution of the climatic situation, especially overthe Pacific, with a view to the operation of its waterresource systems and their future expansion.

Hydrological anomalies are not passive spectators ofthe “surrounding climate” but can contribute to itsshaping. The case of tropical South America is an example of such feedback. We have hypothesized themechanisms by which the hydrology of tropical SouthAmerica plays an important role in constructing the“land–atmosphere bridge” connecting SST in the Pacific to those of the tropical north Atlantic and Caribbean, once an El Niño event develops. In particular,surface pressure, precipitation, temperature, evapotranspiration, soil moisture, and river runoff are highly interconnected in a coherent system. The resulting effectwould produce weakening of the trade wind field thatultimately contributes to the warming of the Caribbeanand the tropical North Atlantic. Dry conditions inducedby El Niño are self-reinforced through the dynamics ofprecipitation recycling and hydrological processes overthe region. Soil moisture over the continent plays a similar role that of sea surface temperature over the oceansin driving the dynamical partition of water and energybudgets at the land–atmosphere and ocean–atmosphereinterfaces, respectively. This paper presents evidence oflong-term persistent behavior of soil moisture anomaliesat interannual timescales over the Amazon, Orinoco, andMagdalena River basins, which suggest that soil moisture over the region is an important component ofENSO. More detailed soil moisture datasets need to becollected and analyzed over the region to confirm theseresults.

Acknowledgments

Support from COLCIENCIAS ofColombia is gratefully acknowledged. Part of this workwas developed while the authors were on leave atCIRES/CSES, University of Colorado, Boulder. Colombian hydrology data were provided by IDEAM and EPMof Colombia. The COADS dataset is from NOAA. Amazon rainfall data are from the EOS-Amazon Project.The soil moisture dataset was provided by J. Schemm(NASA). We thank V. K. Gupta, C. Penland, H. F. Diaz,K. Weickman, K. Trenberth, C. Ropelewski, P. Webster,S. Hastenrath, P. Aceituno, E. A. Davidson, R. Newell,and P. R. Waylen for fruitful discussions and the twoanonymous reviewers for helpful comments.

REFERENCES

  • Aceituno, P., 1988: On the functioning of the Southern Oscillationin the South American sector. Part I: Surface climate. Mon. Wea.Rev.,116, 505–524.

  • ——, 1989: On the functioning of the Southern Oscillation in theSouth American sector. Part II: Upper-air circulation. J. Climate,2, 341–355.

  • Baumgartner, A., and E. Reichel, 1975: The World Water Balance.Elsevier, 179 pp.

  • Brown, B. G., and R. W. Katz, 1991: Use of statistical methods inthe search for teleconnections; Past, present and future. Teleconnections Linking Worldwide Climate Anomalies, R. M.Glantz, R. Katz, and N. Nicholls, Eds., Cambridge UniversityPress, 371–400.

  • Carton, J. A., and B. Huang, 1994: Warm events in the tropicalAtlantic. J. Phys. Oceanogr.,24, 888–903.

  • Chu, P.-S., 1991: Brazil’s climate anomalies and ENSO. Teleconnections Linking Worldwide Climate Anomalies, R. M. Glantz, R.Katz, and N. Nicholls, Eds., Cambridge University Press, 43–71.

  • Covey, D., and S. Hastenrath, 1978: The Pacific El Niño phenomenonand the Atlantic circulation. Mon. Wea. Rev.,106, 1280–1287.

  • Curtis, S., and S. Hastenrath, 1995: Forcing of anomalous sea surfacetemperature evolution in the tropical Atlantic during Pacificwarm events. J. Geophys. Res.,100 (C8), 15835–15847.

  • Diaz, H. F., and V. Markgraf, Eds., 1992: El Niño. Historical andPaleoclimatic Aspects of the Southern Oscillation. CambridgeUniversity Press, 476 pp.

  • Eagleson, P. E., 1994: The evolution of modern hydrology (fromwatershed to continent in 30 years). Adv. Water Res.,17, 3–18.

  • Elthair, E. A. B., and R. Bras, 1994: Precipitation recycling in theAmazon basin. Quart. J. Roy. Meteor. Soc.,120, 861–880.

  • Estoque, M. A., J. Luque, M. Chandeck-Monteza, and J. García, 1985: Effects of El Niño on Panama rainfall. Geofís. Int.,24, 355–381.

  • Frank, N. L., and P. J. Hebert, 1974: Atlantic tropical systems of1973. Mon. Wea. Rev.,102, 290–295.

  • Gessler, R. D., 1995: Precipitation variability over Ecuador associatedwith the El Niño/Southern Oscillation. M.S. thesis, Dept. of Geography, University of Florida, 85 pp. [Available from P. Waylen,Dept. of Geography, University of Florida, Gainesville, FL32611-7315.].

  • Gill, A. E., 1982: Atmosphere–Ocean Dynamics. Academic Press,662 pp.

  • Glantz, M., R. Katz, and N. Nicholls, Eds., 1991: TeleconnectionsLinking Worldwide Climate Anomalies. Cambridge UniversityPress, 535 pp.

  • Gray, W. M., and J. D. Sheaffer, 1991: El Niño and QBO influenceson tropical cyclone activity. Teleconnections Linking WorldwideClimate Anomalies, R. M. Glantz, R. Katz, and N. Nicholls, Eds.,Cambridge University Press, 257–284.

  • Halpert, M. S., and C. F. Ropelewski, 1992: Surface temperaturepatterns associated with the Southern Oscillation. J. Climate,5,577–593.

  • Hastenrath, S., 1976: Variations in low-latitude circulations and extreme climatic events in the tropical Americas. J. Atmos. Sci.,33, 202–215.

  • ——, 1990: Diagnostic and prediction of anomalous river dischargesin northern South America. J. Climate,3, 1080–1096.

  • ——, and M.-C. Wu, 1982: Oscillations of the upper-air circulationand anomalies in the surface climate of the tropics. Arch. Meteor.Geophys. Bioclimatol., Ser. B,31, 1–37.

  • ——, and L. Greischar, 1993: Further work on the prediction ofnortheast Brazil rainfall anomalies. J. Climate,6, 743–758.

  • Hodnett, M. G., M. D. Omaya, J. Tomasella, and A. de O. MarquesFilho, 1996: Comparisons of long-term soil water storage behaviour under pasture and forest in three areas of Amazonia.Amazonian Deforestation and Climate, J. H. C. Gash, C. A.Nobre, J. M. Roberts, and R. L. Victoria, Eds., Wiley and Sons,57–77.

  • Horel, J. D., and J. M. Wallace, 1981: Planetary scale atmosphericphenomena associated with the Southern Oscillation. Mon. Wea.Rev.,109, 813–829.

  • ——, V. E. Kousky, and M. T. Kayano, 1986: Atmospheric conditionsin the Atlantic sector during 1983 and 1984. Nature,322, 248–251.

  • Hsu, H.-H., 1994: Relationship between tropical heating and globalcirculation, Interannual variability. J. Geophys. Res.,99(D5),10473–10489.

  • Jipp, P. H., D. C. Nepstad, D. K. Cassel, and C. R. de Carvalho,1997: Deep soil moisture storage and transpiration in forests andpastures of seasonally-dry Amazonia. Climate Change, in press.

  • Kayano, M. T., V. B. Rao, and A. D. Moura, 1988: Tropical circulations and the associated rainfall anomalies during two contrasting years. J. Climatol.,8, 477–488.

  • Kiladis, G., and H. F. Diaz, 1989: Global climatic anomalies associated with extremes in the Southern Oscillation, J. Climate,2,1069–1090.

  • Kousky, V. E., and M. T. Kayano, 1994: Principal modes of outgoinglongwave radiation and 250-mb circulation for the South American sector. J. Climate,7, 1131–1143.

  • ——, ——, and I. F. A. Cavalcanti, 1984: A review of the SouthernOscillation: Oceanic atmospheric circulation changes and relatedrainfall anomalies. Tellus,36A, 490–504.

  • Lau, K. M., and P. J. Sheu, 1988: Annual cycle, quasi-biennial oscillation, and Southern Oscillation in global precipitation. J.Geophys. Res.,93(D9), 10975–10989.

  • Lau, N.-C., and M. J. Nath, 1994: A modeling study of the relativeroles of tropical and extratropical SST anomalies in the variability of the global atmosphere-ocean system. J. Climate,7,1184–1207.

  • López, M. E., 1966: Cloud seeding trials in the rainy belt of westernColombia. Water Resour. Res.,2, 811–823.

  • ——, and W. E. Howell, 1967: Katabatic winds in the equatorialAndes. J. Atmos. Sci.,24, 29–35.

  • Marengo, J., 1992: Interannual variability of surface climate in theAmazon basin. J. Climatol.,12, 853–863.

  • ——, and S. Hastenrath, 1993: Case studies of extreme climaticevents in the Amazon basin. J. Climate,6, 617–627.

  • Mesa, O. J., and G. Poveda, 1993: The Hurst effect: The scale offluctuation approach. Water Resour. Res.,29, 3995–4002.

  • Nepstad, D. C., and Coauthors, 1994: The role of deep roots in thehydrological and carbon cycles of Amazonian forests and pastures. Nature,372, 666–669.

  • Nigam, S., 1994: On the dynamical basis for the Asian monsoonrainfall–El Niño relationship. J. Climate,7, 1750–1771.

  • Nobre, P., and J. Shukla, 1996: Variations of sea surface temperature,wind stress, and rainfall over the tropical Atlantic and SouthAmerica. J. Climate,9, 2464–2479.

  • Numaguti, A., 1993: Dynamics and energy balance of the Hadleycirculation and the tropical precipitation zones: Significance ofthe distribution of evaporation. J. Atmos. Sci.,50, 1874–1887.

  • Obregón, G. O., and C. A. Nobre, 1990: Principal component analysisof precipitation fields over the Amazon river basin. Climanálise,5, 35–46.

  • Pan, Y. H., and A. H. Oort, 1983: Global climate variations connectedwith the sea surface temperature anomalies in the eastern equatorial Pacific Ocean for the 1958–73 period. Mon. Wea. Rev.,111, 1244–1258.

  • Poveda, G., 1994: Empirical orthogonal functions in the analysis ofthe relationship between mean discharges in Colombia and thePacific and Atlantic Oceans surface temperatures (in Spanish).Proc. XVI Latin-American Congress on Hydraulics and Hydrology, Vol. 4, Santiago, Chile, IAHS, 131–144.

  • ——, and O. J. Mesa, 1993: Methodologies to predict the Colombianhydrology considering the ENSO event (in Spanish). Rev. Atmós.,16, 26–39.

  • ——, and ——, 1995: The relationship between ENSO and the hydrology of tropical South America. The case of Colombia. Proc.Fifteenth Annual American Geophysical Union Hydrology Days,Fort Collins, CO, Hydrology Days Publications, 227–236.[Available from H. Morel-Seytoux, 57 Selby Lane, Atherton,CA 94027-3926.].

  • ——, and ——, 1996: The North Atlantic Oscillation and its influenceon the hydro-climatology of Colombia (in Spanish). Proc. XVIILatin-American Congress on Hydraulics and Hydrology, Vol.II, Guayaquil, Ecuador, IAHR, 343–354.

  • Pulwarty, R. S., 1994: Annual and intrannual variability of convectionover tropical South America. Ph.D. dissertation, University ofColorado, 220 pp. [Available from Dept. of Geography, University of Colorado, Boulder, CO 80305.].

  • ——, and H. F. Diaz, 1993: A study of the seasonal cycle and itsperturbation by ENSO in the tropical Americas. Preprints, FourthInt. Conf. on Southern Hemisphere Meteorology and Oceanography, Hobart, Australia, Amer. Meteor. Soc., 262–263.

  • ——, R. G. Barry, and H. Riehl, 1992: Annual and seasonal patternsof rainfall variability over Venezuela. Erdkunde,46, 273–289.

  • Rao, V. B., and K. Hada, 1990: Characteristics of rainfall over Brazil: Annual variations and connections with the Southern Oscillation.Theor. Appl. Climatol.,42, 81–91.

  • Rasmusson, E. M., 1991: Observational aspects of ENSO cycle teleconnections. Teleconnections Linking Worldwide ClimateAnomalies, Scientifc Basis and Societal Imapcts, M. Glantz, R.W. Katz, and N. Nicholls, Eds. Cambridge University Press, 309–343.

  • ——, and T. H. Carpenter, 1982: Variations in tropical sea surfacetemperature and surface wind fields associated with the SouthernOscillation. Mon. Wea. Rev.,110, 354–384.

  • ——, and K. Mo, 1993: Linkages between 200-mb tropical and extratropical circulation anomalies during the 1986–1989 ENSOcycle. J. Climate,6, 595–616.

  • Richey, J. E., C. Nobre, and C. Deser, 1989: Amazon river dischargeand climate variability: 1903 to 1985. Science,246, 101–103.

  • Riehl, H., and J. S. Malkus, 1958: On the heat balance in the equatorialtrough zone. Geophysica,6, 505–538.

  • Rogers, J. C., 1988: Precipitation variability over the Caribbean andtropical Americas associated with the Southern Oscillation. J.Climate,1, 172–182.

  • Ropelewski, C. F., and M. S. Halpert, 1987: Global and regional scalesprecipitation associated with El Niño–Southern Oscillation.Mon. Wea. Rev.,115, 1606–1626.

  • Schemm, J., S. Schubert, J. Terry, and S. Bloom, 1992: Estimates ofmonthly mean soil moisture for 1979–1989. NASA Tech. Memo.104571, Goddard Space Flight Center, Greenbelt, MD, 252 pp.[Available from S. Schubert, Goddard Space Flight Center,Greenbelt, MD 20771.].

  • Shuttleworth, W. J., 1988: Evaporation from Amazonian rainforest.Philos. Trans. Roy. Soc. London, Ser. B,233, 321–346.

  • Silva Dias, P. L., J. P. Bonatti, and V. E. Kousky, 1987: Diurnallyforced tropical tropospheric circulation over South America.Mon. Wea Rev.,115, 1465–1478.

  • Slutz, R. J., S. J. Lubker, J. D. Hiscox, S. D. Woodruff, R. L. Jenne,D. H. Joseph, P. M. Steurer, and J. D. Elms, 1985: ComprehensiveOcean-Atmosphere Data Set, release 1. NOAA Tech. Note, 255pp. [Available from University of Colorado/National Oceanicand Atmospheric Administration, Cooperative Institute for Research in Environmental Sciences, Boulder, CO 80305.].

  • Trenberth, K., 1991: General characteristics of El Niño–Southern Oscillation. Teleconnections Linking Worldwide Climate Anomalies, R. M. Glantz, R. Katz, and N. Nicholls, Eds., CambridgeUniversity Press, 13–42.

  • van Loon, H., and R. A. Madden, 1981: The Southern Oscillation,I, Global associations with pressure and temperature in northernwinter. Mon. Wea. Rev.,109, 1150–1162.

  • Vanmarcke, E., 1988: Random Fields: Analysis and Synthesis. TheMIT Press, 382 pp.

  • Velasco, I., and M. Fritsch, 1987: Mesoscale convective complexesin the Americas. J. Geophys. Res.,92(D8), 9591–9613.

  • Vörösmarty, C. J., C. J. Willmott, B. J. Choudhury, A. L. Schloss,T. K. Stearns, S. M. Robeson, and T. J. Dorman, 1996: Analyzingthe discharge regime of a large tropical river through remotesensing, ground-based climatic data, and modeling. Water Resour. Res.,32, 3137–3150.

  • Wagner, R. C., 1996: Mechanisms controlling variability of the interhemispheric sea surface temperature gradient in the tropicalAtlantic. J. Climate,9, 2010–2019.

  • Waylen, P. R., C. N. Caviedes, and M. E. Quesada, 1996: Interannualvariability of monthly precipitation in Costa Rica. J. Climate,9, 2606–2613.

  • Yasunari, T., 1987: Global structure of the El Niño/Southern Oscillation Part I. El Niño composites. J. Meteor. Soc. Japan,65,67–79.

  • Zebiak, S. E., 1993: Air–sea interaction in the equatorial Atlanticregion. J. Climate,6, 1567–1586.

  • Zeng, N., R. E. Dickinson, and X. Zeng, 1996: Climatic impact ofAmazon deforestation—A mechanistic model study. J. Climate,9, 859–883.

  • Zhang, H., K., McGuffie, and A. Henderson-Sellers, 1996: Impactsof tropical deforestation. Part II: The role of large-scale dynamics. J. Climate,9, 2498–2521.

Fig. 1.
Fig. 1.

Time evolution of the SOI (dotted line), the monthly discharges of the Cauca River,Salvajina, Colombia (solid thick line), and average precipitation in the Amazon River basin.Data have been low-pass (12-month) filtered.

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 2.
Fig. 2.

Monthly anomalies of the Magdalena River at Puerto Berrío,Colombia, with respect to the multiannual mean for El Niño (triangles) and La Niña (squares) years (top). Ratio of the average monthlydischarges for El Niño (triangles) and La Niña (squares) to the multiannual averages (bottom).

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 3.
Fig. 3.

Behavior of the cross correlations between the SOI and thestreamflows of four rivers in Colombia. From west to east: La Vieja(Cartago, Valle del Cauca, 4°46′N, 75°54′W), Luisa (Pavo Real, Tolima, 4°13′N, 75°12′W), Sumapaz (El Profundo, Cundinamarca4°00′N, 74°30′W), and Chivor (Ubalá, Boyacá, 4°47′N, 73°09′W).Negative lags correspond to the SOI leading the hydrology. Noticethat the peaks of the cross correlations occur later as the stations arefarther east. The correlations are statistically significant at the 99.9%level.

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 4.
Fig. 4.

Time–longitude diagram depicting rainfall anomalies in the whole Amazon basin during the 1972–92 period. Anomalies have beensmoothed using a low-pass 12-month digital filter.

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 5.
Fig. 5.

Location in Colombia of the 88 rainfall stations and the 55river discharge stations included in the separated EOF analyses.

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 6.
Fig. 6.

The first two EOFs for the Colombian monthly rainfall (3-month running means), which explain 30.7% and 6.5% of the totalvariance, respectively.

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 7.
Fig. 7.

(a) Time series of the first principal component (PC) for the Colombian streamflowriver discharges. Monthly values have been smoothed as 3-month running means. (b) Fast Fouriertransform of the first PC. Frequency is relative to the total record length (396 months), and thecorresponding period (months) is estimated as 396/relative frequency.

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 8.
Fig. 8.

Isocorrelations (%) between sea surface temperatures at each location and the firstprincipal component of the Colombian monthly streamflows (3-month running averages). Noticethe high values of the correlation coefficients at the regions of the central Pacific Ocean andthe region of the Indian monsoon.

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 9.
Fig. 9.

Lagged isocorrelations (%) between the first principal component of the Colombian river discharges (3-month running means) andsea surface temperatures at each location. Streamflows are leading the Atlantic SSTs from 1 to 6 months. Notice that the region of largercorrelations is traveling eastward and gaining areal extent with time, and peaks at 4–5 months.

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 10.
Fig. 10.

Lagged isocorrelations (%) between the Southern Oscillation index (SOI) and sea surface temperatures at each location. SOI isleading the SSTs from 1 to 6 months. Notice that the correlations are lower than those shown in Fig. 9.

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 11.
Fig. 11.

Time evolution of monthly average ground wetness anomalies for the combinedAmazon, Orinoco, and Magdalena River basins during 1979–92. Monthly averages have beensmoothed using a 12-month low-pass filter (thick line).

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Fig. 12.
Fig. 12.

Time evolution of standardized anomalies for the meansea level pressure gradient between the Azores Islands (25°–35°W,30°–40°N) and northern South America (50°–70°W, 3°S–10°N), andprecipitation over the latter region, for the period 1982–94. Data fromthe NCEP–NCAR reanalysis project.

Citation: Journal of Climate 10, 10; 10.1175/1520-0442(1997)010<2690:FBHPIT>2.0.CO;2

Table 1.

Lag-zero correlation coefficients (%) between rainfall in Colombia and climatic indices over the Pacific Ocean (12-monthrunning averages).

Table 1.
Save
  • Aceituno, P., 1988: On the functioning of the Southern Oscillationin the South American sector. Part I: Surface climate. Mon. Wea.Rev.,116, 505–524.

  • ——, 1989: On the functioning of the Southern Oscillation in theSouth American sector. Part II: Upper-air circulation. J. Climate,2, 341–355.

  • Baumgartner, A., and E. Reichel, 1975: The World Water Balance.Elsevier, 179 pp.

  • Brown, B. G., and R. W. Katz, 1991: Use of statistical methods inthe search for teleconnections; Past, present and future. Teleconnections Linking Worldwide Climate Anomalies, R. M.Glantz, R. Katz, and N. Nicholls, Eds., Cambridge UniversityPress, 371–400.

  • Carton, J. A., and B. Huang, 1994: Warm events in the tropicalAtlantic. J. Phys. Oceanogr.,24, 888–903.

  • Chu, P.-S., 1991: Brazil’s climate anomalies and ENSO. Teleconnections Linking Worldwide Climate Anomalies, R. M. Glantz, R.Katz, and N. Nicholls, Eds., Cambridge University Press, 43–71.

  • Covey, D., and S. Hastenrath, 1978: The Pacific El Niño phenomenonand the Atlantic circulation. Mon. Wea. Rev.,106, 1280–1287.

  • Curtis, S., and S. Hastenrath, 1995: Forcing of anomalous sea surfacetemperature evolution in the tropical Atlantic during Pacificwarm events. J. Geophys. Res.,100 (C8), 15835–15847.

  • Diaz, H. F., and V. Markgraf, Eds., 1992: El Niño. Historical andPaleoclimatic Aspects of the Southern Oscillation. CambridgeUniversity Press, 476 pp.

  • Eagleson, P. E., 1994: The evolution of modern hydrology (fromwatershed to continent in 30 years). Adv. Water Res.,17, 3–18.

  • Elthair, E. A. B., and R. Bras, 1994: Precipitation recycling in theAmazon basin. Quart. J. Roy. Meteor. Soc.,120, 861–880.

  • Estoque, M. A., J. Luque, M. Chandeck-Monteza, and J. García, 1985: Effects of El Niño on Panama rainfall. Geofís. Int.,24, 355–381.

  • Frank, N. L., and P. J. Hebert, 1974: Atlantic tropical systems of1973. Mon. Wea. Rev.,102, 290–295.

  • Gessler, R. D., 1995: Precipitation variability over Ecuador associatedwith the El Niño/Southern Oscillation. M.S. thesis, Dept. of Geography, University of Florida, 85 pp. [Available from P. Waylen,Dept. of Geography, University of Florida, Gainesville, FL32611-7315.].

  • Gill, A. E., 1982: Atmosphere–Ocean Dynamics. Academic Press,662 pp.

  • Glantz, M., R. Katz, and N. Nicholls, Eds., 1991: TeleconnectionsLinking Worldwide Climate Anomalies. Cambridge UniversityPress, 535 pp.

  • Gray, W. M., and J. D. Sheaffer, 1991: El Niño and QBO influenceson tropical cyclone activity. Teleconnections Linking WorldwideClimate Anomalies, R. M. Glantz, R. Katz, and N. Nicholls, Eds.,Cambridge University Press, 257–284.

  • Halpert, M. S., and C. F. Ropelewski, 1992: Surface temperaturepatterns associated with the Southern Oscillation. J. Climate,5,577–593.

  • Hastenrath, S., 1976: Variations in low-latitude circulations and extreme climatic events in the tropical Americas. J. Atmos. Sci.,33, 202–215.

  • ——, 1990: Diagnostic and prediction of anomalous river dischargesin northern South America. J. Climate,3, 1080–1096.

  • ——, and M.-C. Wu, 1982: Oscillations of the upper-air circulationand anomalies in the surface climate of the tropics. Arch. Meteor.Geophys. Bioclimatol., Ser. B,31, 1–37.

  • ——, and L. Greischar, 1993: Further work on the prediction ofnortheast Brazil rainfall anomalies. J. Climate,6, 743–758.

  • Hodnett, M. G., M. D. Omaya, J. Tomasella, and A. de O. MarquesFilho, 1996: Comparisons of long-term soil water storage behaviour under pasture and forest in three areas of Amazonia.Amazonian Deforestation and Climate, J. H. C. Gash, C. A.Nobre, J. M. Roberts, and R. L. Victoria, Eds., Wiley and Sons,57–77.

  • Horel, J. D., and J. M. Wallace, 1981: Planetary scale atmosphericphenomena associated with the Southern Oscillation. Mon. Wea.Rev.,109, 813–829.

  • ——, V. E. Kousky, and M. T. Kayano, 1986: Atmospheric conditionsin the Atlantic sector during 1983 and 1984. Nature,322, 248–251.

  • Hsu, H.-H., 1994: Relationship between tropical heating and globalcirculation, Interannual variability. J. Geophys. Res.,99(D5),10473–10489.

  • Jipp, P. H., D. C. Nepstad, D. K. Cassel, and C. R. de Carvalho,1997: Deep soil moisture storage and transpiration in forests andpastures of seasonally-dry Amazonia. Climate Change, in press.

  • Kayano, M. T., V. B. Rao, and A. D. Moura, 1988: Tropical circulations and the associated rainfall anomalies during two contrasting years. J. Climatol.,8, 477–488.

  • Kiladis, G., and H. F. Diaz, 1989: Global climatic anomalies associated with extremes in the Southern Oscillation, J. Climate,2,1069–1090.

  • Kousky, V. E., and M. T. Kayano, 1994: Principal modes of outgoinglongwave radiation and 250-mb circulation for the South American sector. J. Climate,7, 1131–1143.

  • ——, ——, and I. F. A. Cavalcanti, 1984: A review of the SouthernOscillation: Oceanic atmospheric circulation changes and relatedrainfall anomalies. Tellus,36A, 490–504.

  • Lau, K. M., and P. J. Sheu, 1988: Annual cycle, quasi-biennial oscillation, and Southern Oscillation in global precipitation. J.Geophys. Res.,93(D9), 10975–10989.

  • Lau, N.-C., and M. J. Nath, 1994: A modeling study of the relativeroles of tropical and extratropical SST anomalies in the variability of the global atmosphere-ocean system. J. Climate,7,1184–1207.

  • López, M. E., 1966: Cloud seeding trials in the rainy belt of westernColombia. Water Resour. Res.,2, 811–823.

  • ——, and W. E. Howell, 1967: Katabatic winds in the equatorialAndes. J. Atmos. Sci.,24, 29–35.

  • Marengo, J., 1992: Interannual variability of surface climate in theAmazon basin. J. Climatol.,12, 853–863.

  • ——, and S. Hastenrath, 1993: Case studies of extreme climaticevents in the Amazon basin. J. Climate,6, 617–627.

  • Mesa, O. J., and G. Poveda, 1993: The Hurst effect: The scale offluctuation approach. Water Resour. Res.,29, 3995–4002.

  • Nepstad, D. C., and Coauthors, 1994: The role of deep roots in thehydrological and carbon cycles of Amazonian forests and pastures. Nature,372, 666–669.

  • Nigam, S., 1994: On the dynamical basis for the Asian monsoonrainfall–El Niño relationship. J. Climate,7, 1750–1771.

  • Nobre, P., and J. Shukla, 1996: Variations of sea surface temperature,wind stress, and rainfall over the tropical Atlantic and SouthAmerica. J. Climate,9, 2464–2479.

  • Numaguti, A., 1993: Dynamics and energy balance of the Hadleycirculation and the tropical precipitation zones: Significance ofthe distribution of evaporation. J. Atmos. Sci.,50, 1874–1887.

  • Obregón, G. O., and C. A. Nobre, 1990: Principal component analysisof precipitation fields over the Amazon river basin. Climanálise,5, 35–46.

  • Pan, Y. H., and A. H. Oort, 1983: Global climate variations connectedwith the sea surface temperature anomalies in the eastern equatorial Pacific Ocean for the 1958–73 period. Mon. Wea. Rev.,111, 1244–1258.

  • Poveda, G., 1994: Empirical orthogonal functions in the analysis ofthe relationship between mean discharges in Colombia and thePacific and Atlantic Oceans surface temperatures (in Spanish).Proc. XVI Latin-American Congress on Hydraulics and Hydrology, Vol. 4, Santiago, Chile, IAHS, 131–144.

  • ——, and O. J. Mesa, 1993: Methodologies to predict the Colombianhydrology considering the ENSO event (in Spanish). Rev. Atmós.,16, 26–39.

  • ——, and ——, 1995: The relationship between ENSO and the hydrology of tropical South America. The case of Colombia. Proc.Fifteenth Annual American Geophysical Union Hydrology Days,Fort Collins, CO, Hydrology Days Publications, 227–236.[Available from H. Morel-Seytoux, 57 Selby Lane, Atherton,CA 94027-3926.].

  • ——, and ——, 1996: The North Atlantic Oscillation and its influenceon the hydro-climatology of Colombia (in Spanish). Proc. XVIILatin-American Congress on Hydraulics and Hydrology, Vol.II, Guayaquil, Ecuador, IAHR, 343–354.

  • Pulwarty, R. S., 1994: Annual and intrannual variability of convectionover tropical South America. Ph.D. dissertation, University ofColorado, 220 pp. [Available from Dept. of Geography, University of Colorado, Boulder, CO 80305.].

  • ——, and H. F. Diaz, 1993: A study of the seasonal cycle and itsperturbation by ENSO in the tropical Americas. Preprints, FourthInt. Conf. on Southern Hemisphere Meteorology and Oceanography, Hobart, Australia, Amer. Meteor. Soc., 262–263.

  • ——, R. G. Barry, and H. Riehl, 1992: Annual and seasonal patternsof rainfall variability over Venezuela. Erdkunde,46, 273–289.

  • Rao, V. B., and K. Hada, 1990: Characteristics of rainfall over Brazil: Annual variations and connections with the Southern Oscillation.Theor. Appl. Climatol.,42, 81–91.

  • Rasmusson, E. M., 1991: Observational aspects of ENSO cycle teleconnections. Teleconnections Linking Worldwide ClimateAnomalies, Scientifc Basis and Societal Imapcts, M. Glantz, R.W. Katz, and N. Nicholls, Eds. Cambridge University Press, 309–343.

  • ——, and T. H. Carpenter, 1982: Variations in tropical sea surfacetemperature and surface wind fields associated with the SouthernOscillation. Mon. Wea. Rev.,110, 354–384.

  • ——, and K. Mo, 1993: Linkages between 200-mb tropical and extratropical circulation anomalies during the 1986–1989 ENSOcycle. J. Climate,6, 595–616.

  • Richey, J. E., C. Nobre, and C. Deser, 1989: Amazon river dischargeand climate variability: 1903 to 1985. Science,246, 101–103.

  • Riehl, H., and J. S. Malkus, 1958: On the heat balance in the equatorialtrough zone. Geophysica,6, 505–538.

  • Rogers, J. C., 1988: Precipitation variability over the Caribbean andtropical Americas associated with the Southern Oscillation. J.Climate,1, 172–182.

  • Ropelewski, C. F., and M. S. Halpert, 1987: Global and regional scalesprecipitation associated with El Niño–Southern Oscillation.Mon. Wea. Rev.,115, 1606–1626.

  • Schemm, J., S. Schubert, J. Terry, and S. Bloom, 1992: Estimates ofmonthly mean soil moisture for 1979–1989. NASA Tech. Memo.104571, Goddard Space Flight Center, Greenbelt, MD, 252 pp.[Available from S. Schubert, Goddard Space Flight Center,Greenbelt, MD 20771.].

  • Shuttleworth, W. J., 1988: Evaporation from Amazonian rainforest.Philos. Trans. Roy. Soc. London, Ser. B,233, 321–346.

  • Silva Dias, P. L., J. P. Bonatti, and V. E. Kousky, 1987: Diurnallyforced tropical tropospheric circulation over South America.Mon. Wea Rev.,115, 1465–1478.

  • Slutz, R. J., S. J. Lubker, J. D. Hiscox, S. D. Woodruff, R. L. Jenne,D. H. Joseph, P. M. Steurer, and J. D. Elms, 1985: ComprehensiveOcean-Atmosphere Data Set, release 1. NOAA Tech. Note, 255pp. [Available from University of Colorado/National Oceanicand Atmospheric Administration, Cooperative Institute for Research in Environmental Sciences, Boulder, CO 80305.].

  • Trenberth, K., 1991: General characteristics of El Niño–Southern Oscillation. Teleconnections Linking Worldwide Climate Anomalies, R. M. Glantz, R. Katz, and N. Nicholls, Eds., CambridgeUniversity Press, 13–42.

  • van Loon, H., and R. A. Madden, 1981: The Southern Oscillation,I, Global associations with pressure and temperature in northernwinter. Mon. Wea. Rev.,109, 1150–1162.

  • Vanmarcke, E., 1988: Random Fields: Analysis and Synthesis. TheMIT Press, 382 pp.

  • Velasco, I., and M. Fritsch, 1987: Mesoscale convective complexesin the Americas. J. Geophys. Res.,92(D8), 9591–9613.

  • Vörösmarty, C. J., C. J. Willmott, B. J. Choudhury, A. L. Schloss,T. K. Stearns, S. M. Robeson, and T. J. Dorman, 1996: Analyzingthe discharge regime of a large tropical river through remotesensing, ground-based climatic data, and modeling. Water Resour. Res.,32, 3137–3150.

  • Wagner, R. C., 1996: Mechanisms controlling variability of the interhemispheric sea surface temperature gradient in the tropicalAtlantic. J. Climate,9, 2010–2019.

  • Waylen, P. R., C. N. Caviedes, and M. E. Quesada, 1996: Interannualvariability of monthly precipitation in Costa Rica. J. Climate,9, 2606–2613.

  • Yasunari, T., 1987: Global structure of the El Niño/Southern Oscillation Part I. El Niño composites. J. Meteor. Soc. Japan,65,67–79.

  • Zebiak, S. E., 1993: Air–sea interaction in the equatorial Atlanticregion. J. Climate,6, 1567–1586.

  • Zeng, N., R. E. Dickinson, and X. Zeng, 1996: Climatic impact ofAmazon deforestation—A mechanistic model study. J. Climate,9, 859–883.

  • Zhang, H., K., McGuffie, and A. Henderson-Sellers, 1996: Impactsof tropical deforestation. Part II: The role of large-scale dynamics. J. Climate,9, 2498–2521.

  • Fig. 1.

    Time evolution of the SOI (dotted line), the monthly discharges of the Cauca River,Salvajina, Colombia (solid thick line), and average precipitation in the Amazon River basin.Data have been low-pass (12-month) filtered.

  • Fig. 2.

    Monthly anomalies of the Magdalena River at Puerto Berrío,Colombia, with respect to the multiannual mean for El Niño (triangles) and La Niña (squares) years (top). Ratio of the average monthlydischarges for El Niño (triangles) and La Niña (squares) to the multiannual averages (bottom).

  • Fig. 3.

    Behavior of the cross correlations between the SOI and thestreamflows of four rivers in Colombia. From west to east: La Vieja(Cartago, Valle del Cauca, 4°46′N, 75°54′W), Luisa (Pavo Real, Tolima, 4°13′N, 75°12′W), Sumapaz (El Profundo, Cundinamarca4°00′N, 74°30′W), and Chivor (Ubalá, Boyacá, 4°47′N, 73°09′W).Negative lags correspond to the SOI leading the hydrology. Noticethat the peaks of the cross correlations occur later as the stations arefarther east. The correlations are statistically significant at the 99.9%level.

  • Fig. 4.

    Time–longitude diagram depicting rainfall anomalies in the whole Amazon basin during the 1972–92 period. Anomalies have beensmoothed using a low-pass 12-month digital filter.

  • Fig. 5.

    Location in Colombia of the 88 rainfall stations and the 55river discharge stations included in the separated EOF analyses.

  • Fig. 6.

    The first two EOFs for the Colombian monthly rainfall (3-month running means), which explain 30.7% and 6.5% of the totalvariance, respectively.

  • Fig. 7.

    (a) Time series of the first principal component (PC) for the Colombian streamflowriver discharges. Monthly values have been smoothed as 3-month running means. (b) Fast Fouriertransform of the first PC. Frequency is relative to the total record length (396 months), and thecorresponding period (months) is estimated as 396/relative frequency.

  • Fig. 8.

    Isocorrelations (%) between sea surface temperatures at each location and the firstprincipal component of the Colombian monthly streamflows (3-month running averages). Noticethe high values of the correlation coefficients at the regions of the central Pacific Ocean andthe region of the Indian monsoon.

  • Fig. 9.

    Lagged isocorrelations (%) between the first principal component of the Colombian river discharges (3-month running means) andsea surface temperatures at each location. Streamflows are leading the Atlantic SSTs from 1 to 6 months. Notice that the region of largercorrelations is traveling eastward and gaining areal extent with time, and peaks at 4–5 months.

  • Fig. 10.

    Lagged isocorrelations (%) between the Southern Oscillation index (SOI) and sea surface temperatures at each location. SOI isleading the SSTs from 1 to 6 months. Notice that the correlations are lower than those shown in Fig. 9.

  • Fig. 11.

    Time evolution of monthly average ground wetness anomalies for the combinedAmazon, Orinoco, and Magdalena River basins during 1979–92. Monthly averages have beensmoothed using a 12-month low-pass filter (thick line).

  • Fig. 12.

    Time evolution of standardized anomalies for the meansea level pressure gradient between the Azores Islands (25°–35°W,30°–40°N) and northern South America (50°–70°W, 3°S–10°N), andprecipitation over the latter region, for the period 1982–94. Data fromthe NCEP–NCAR reanalysis project.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 957 222 16
PDF Downloads 585 111 10