South American Rainfall Dipole (SAD)

• South American Rainfall Dipole (SAD) is a recurring dipolar pattern where one region experiences anomalous wetness and another severe dryness, driven by varying oceanic forcings.
• The Amazonian formulation distinguishes Northeastern and Southwestern Amazon responses, linking NA dryness to El Niño and SA dryness to tropical North Atlantic warming.
• The austral-summer formulation reveals SACZ–SESA rainfall opposition modulated by extreme MJO events and circulation changes, offering insights for drought monitoring.

Searching arXiv for the specified papers and related South American Rainfall Dipole literature. The South American Rainfall Dipole (SAD) denotes a dipolar organization of precipitation anomalies over South America in which one sector is anomalously wet while another is anomalously dry. In the literature considered here, the term is used in two related but not identical senses. One formulation identifies a dipole within Amazonia, contrasting Northeastern Amazon (NA) and Southwestern Amazon (SA), with NA dryness linked primarily to El Niño and SA dryness linked primarily to tropical North Atlantic warming (Zou et al., 2015). A second formulation, analyzed for austral summer, describes a dipolar rainfall pattern with opposite-signed anomalies between the South Atlantic Convergence Zone (SACZ; tropical/eastern Brazil) and southeastern South America (SESA; southern Brazil–Uruguay–northern Argentina), and shows that extreme Madden–Julian Oscillation (MJO) events modulate its intensity, spatial extent, and phase evolution (Minjares et al., 27 Jul 2025). Taken together, these studies present SAD less as a single fixed index than as a recurring regional mode of opposite-signed rainfall anomalies whose structure depends on spatial domain, timescale, and forcing.

1. Definitions and regional formulations

The term SAD is not used with a single universal geometry in the cited literature. In the Amazonian formulation, the dipole is expressed between NA and SA, where the analyzed domains are NA: 65°W–48°W, 5°N–10°S, and SA: 75°W–50°W, 15°S–4°S. In that framework, the central result is a dipolar rainfall response in which very dry periods in NA are associated with El Niño conditions, whereas very dry periods in SA are associated with anomalous warming in the tropical North Atlantic (Zou et al., 2015).

In the austral-summer intraseasonal formulation, SAD is described qualitatively as a dipolar rainfall pattern with opposite-signed anomalies between the SACZ and SESA. The paper explicitly states that during phases 8 and 1 (3 and 4) of MJO the SACZ (SESA) is in a wet phase and conversely for SESA (SACZ). In this usage, SAD is not represented by a formal scalar index, fixed geographic boxes, or a closed-form metric; it is diagnosed from composited anomaly patterns, primarily outgoing longwave radiation (OLR) as a rainfall proxy, together with upper-level circulation anomalies (Minjares et al., 27 Jul 2025).

A concise comparison is therefore useful.

Formulation	Spatial contrast	Primary diagnostic
Amazonian SAD	NA versus SA	Daily precipitation anomaly and drought period length
Austral-summer SAD	SACZ versus SESA	OLR anomaly patterns and 200 hPa circulation composites

This difference in usage is a recurrent source of ambiguity. A common misconception is to treat SAD as a single invariant mode with a single index. The available evidence instead indicates that the term has been applied to at least two domain-specific dipolar structures: one centered on Amazonian drought variability, and another centered on SACZ–SESA rainfall opposition during austral summer.

2. Amazonian dipole structure and drought attribution

In the Amazonian analysis, SAD emerges from the contrasting behavior of NA and SA during the severe droughts of 2005 and 2010. The study proposes a daily drought period length (DPL) measure to assess the temporal difference between the two events and shows that the two subregions respond differently to sea surface temperature (SST) variability. The Pacific and Atlantic oceans have different roles on precipitation patterns in Amazonia: very dry periods in NA are influenced by El Niño events, while very dry periods in SA are affected by anomalously warming SST in the North Atlantic (Zou et al., 2015).

The 2005 drought is characterized as an SA-focused event. In SA, an extreme very dry period begins on day 173 of 2005, corresponding to 22 June, and reaches over 40 days of continuous negative precipitation anomalies. The same event is described as involving about 60 days of rainfall less than 1 mm/day in SA. By contrast, NA shows no very dry periods exceeding 20 days; only mild or dry conditions are reported. Niño3.4 anomalies remained near the 0.4°C threshold with small fluctuations and did not satisfy the ENSO event criterion long enough to explain the drought, whereas tropical North Atlantic SST anomalies were significantly above 0.4°C during the key period. The attribution given is explicit: “The drought 2005 hit SA, which is caused by the North Atlantic only.”

The 2010 drought is presented as a two-phase, sequential dipole. Phase (i) begins in NA on 1 August 2009, day 213 of 2009, during El Niño conditions. NA experiences a very dry period exceeding 40 days of negative precipitation anomalies, and this dryness extends into the first half of 2010, with multiple dry spells of up to about 20 days. During this first phase, SA is described as having “plenty of rain,” with dry periods much less than 10 days, which exemplifies opposite-signed dipolar behavior. Phase (ii) begins in SA around August 2010, when SA registers at least two very dry events exceeding 35 days, associated with anomalous warming in the tropical North Atlantic. The result is a much larger spatial and temporal coverage of rainfall shortage than in 2005.

These two events illustrate two distinct realizations of the Amazonian SAD. In 2005 the dipole is essentially SA-dominant, with SA dry and NA near-normal. In 2010 the dipole is sequential, first with NA dry during El Niño and then with SA dry during tropical North Atlantic warming. This suggests that the dipole need not be strictly simultaneous; it may also manifest as temporally ordered regional responses to different oceanic forcings.

3. Diagnostic frameworks and quantitative definitions

The Amazonian study formalizes dry-spell persistence with daily precipitation anomalies and DPL. For each subregion, the precipitation anomaly is defined relative to the climatological daily mean over 1961–2000,

$$A(t) = P(t) - \text{climatological daily mean over 1961–2000 for the corresponding calendar day}.$$

The drought period length is the waiting time from day $t$ until the next day with a non-negative anomaly,

$$DPL(t) = \tau,\qquad \tau = \min\{\tau: A(t+\tau)\ge 0,\ \tau \in [0,\infty)\}.$$

A dry spell is a contiguous run of days with $A(t)<0$. Over the full analysis period, the maximum observed DPL is 42 days, which yields the category thresholds mild dry: $DPL \le 14$ days, dry: $14 < DPL \le 29$ days, and very dry: $DPL > 29$ days. The anomalies are neither normalized nor standardized (Zou et al., 2015).

The data architecture is correspondingly explicit. Precipitation is taken from the Princeton Global Forcings dataset at 0.5° resolution for 1948–2010, with daily time series area-averaged over the NA and SA domains. SST is taken from NOAA High Resolution SST products, with area averages over Niño3.4, tropical North Atlantic (n-Atlan), and tropical South Atlantic (s-Atlan) domains. For ENSO identification, the 5-month running mean of Niño3.4 SST anomalies must exceed 0.4°C for at least 6 consecutive months:

$I_{N34}(t)=\frac{1}{5}\sum_{k=0}^{4} S'(t-k).$

The austral-summer MJO study uses a different diagnostic architecture. The MJO is represented by the Real-time Multivariate MJO index with amplitude and phase

$$A=\sqrt{RMM1^2+RMM2^2},\qquad \phi=\arctan(RMM2/RMM1).$$

Active MJO days satisfy $A \ge 1$, and an event is any contiguous period with $t$0. Event size is defined as

$t$1

A power-law fit to the tail of the event-size distribution is used to define extreme events, and the operational threshold is size $t$2, with weak events defined as size $t$3. The study composites OLR anomalies, 200 hPa eddy streamfunction $t$4, and 200 hPa velocity potential $t$5, with divergent wind given by

$t$6

Unlike the Amazonian DPL framework, this analysis does not define a formal SAD index, does not provide geographic boxes for SAD itself, and does not provide precipitation anomalies in mm day$t$7 (Minjares et al., 27 Jul 2025).

The methodological contrast is substantive. The DPL framework is event-centric and persistence-based, tailored to onset and duration of negative rainfall anomalies. The MJO framework is phase-composite and circulation-centric, tailored to spatial organization and evolution of dipolar anomaly patterns on intraseasonal time scales.

4. Oceanic forcing, lags, and dynamical interpretation

In the Amazonian SAD, the forcing asymmetry between Pacific and Atlantic sectors is central. Very dry periods in NA are associated with El Niño conditions, identified by Niño3.4 SST anomalies exceeding 0.4°C, and the rainfall response exhibits a qualitative lag of 2–3 months. Very dry periods in SA are associated with anomalous warming in the tropical North Atlantic, with an approximately 1-month lag. Scatter plots of dry and very dry days against SST anomalies support this asymmetry: NA dry days cluster with positive Niño3.4 anomalies, whereas SA dry days cluster with tropical North Atlantic warming; tropical South Atlantic cooling does not play a direct role in very dry days, but helps create the Atlantic meridional gradient during 2005 and 2010 (Zou et al., 2015).

The associated physical interpretation is framed in terms of large-scale overturning and ITCZ displacement. For ENSO, the paper invokes anomalous Walker circulation to connect Pacific warming to dry conditions over Amazonia, with the effect “limited to the lower latitude of northeast,” that is, NA. For tropical North Atlantic warming, the interpretation is a meridional Hadley circulation response: anomalously high TNA SSTs establish upward motion over the North Atlantic and subsidence over Amazonia, suppressing precipitation. When anomalous TNA warming is coupled with anomalous TSA cooling, the north–south Atlantic temperature gradient is enhanced, displacing the ITCZ northward and reducing moisture convergence over SA.

The 2010 event illustrates overlap between Pacific and Atlantic influences. The paper notes an overlapped window in 2010 when El Niño and TNA warming both exceed 0.4°C, together with a rapid transition from El Niño to La Niña that favors a large northward Atlantic temperature gradient. A plausible implication is that basin-wide drought severity in 2010 arose not from a single forcing pathway but from the temporal concatenation of Pacific-forced NA dryness and Atlantic-forced SA dryness.

These mechanisms are interpretive rather than fully diagnosed in the paper. No direct diagnostics of vertical velocity, wind fields, or moisture flux convergence are presented. The mechanistic account is therefore conceptual and literature-supported, rather than established by a closed dynamical budget.

5. Intraseasonal SAD modulation by extreme MJO events

On intraseasonal scales, the austral-summer study shows that extreme MJO events influence the SAD and produce more intense rainfall anomalies of larger spatial extent than weak events. In this formulation, the canonical SAD during MJO initiation in phases 2–3 consists of wet SESA and dry SACZ. Extreme events produce strong negative OLR anomalies over northern South America and SESA, with positive OLR over eastern Brazil/SACZ. The associated 200 hPa eddy streamfunction shows a coherent upper-level trough pattern with cyclonic–anticyclonic anomalies over subtropical South America, favoring dynamic lift over SESA and subsidence over SACZ. Weak events produce a similar but weaker and more localized pattern (Minjares et al., 27 Jul 2025).

As the MJO evolves from phases 2–3 into phases 4–5, the SAD intensifies and expands spatially for extreme events. The subtropical anticyclonic component of $t$8 weakens but remains influential for SESA rainfall, while velocity potential and divergent winds show strong upper-level convergence over equatorial South America, reinforcing subsidence over the SACZ. By phases 6–7, the dipole largely disappears in the composites, and negative OLR anomalies peak instead over the central Andes and the northern Chile–Bolivia–Peru sector. By phases 8–1, the SAD reverses sign, with wet SACZ and dry SESA, consistent with a reversal in the circulation anomalies.

The study also analyzes events initiating in phases 6–7. In those cases, extreme events show stronger and farther eastward convection and circulation anomalies than weak events. The most prominent SACZ enhancement occurs in phases 8–1, especially for extreme events, while later phases exhibit fewer cases and greater noise. The southeastward displacement of enhanced and suppressed convection during extreme events, relative to weak events, is identified as a key structural difference because it couples tropical convective anomalies more effectively to the subtropical South American Rossby wave response, producing a stronger and more expansive SAD.

A substantial difference between extreme and weak events lies in propagation and persistence. More than 68% of extreme events that reach phases 4–5 propagate beyond the Maritime Continent, whereas more than 50% of weak events starting in any phase are lost before reaching the next phase. In the 2–3 initiation pathway, only 9 weak events, totaling 29 days, reach phases 6–7. This helps explain why the dipole structure in weak composites is often faint, spatially confined, or not robust.

6. Predictability, ambiguity, and limitations

Both studies present the SAD as potentially useful for monitoring and early warning, but the predictability claims are tightly conditioned by diagnostic choice and by unresolved uncertainties. In the Amazonian study, monitoring Niño3.4 anomalies and TNA/TSA gradients provides lead information: NA dryness is expected about 2–3 months after strong El Niño onset, whereas SA dryness is expected about 1 month after significant TNA warming, especially if TSA cools simultaneously. Overlapping ENSO and TNA warming, as in 2010, can produce sequential phases and broader spatial coverage of drought risk (Zou et al., 2015).

In the MJO study, the operational monitoring variables are MJO amplitude, phase, and event size. Active days are defined by $t$9, and extreme-like behavior is associated with event size $$DPL(t) = \tau,\qquad \tau = \min\{\tau: A(t+\tau)\ge 0,\ \tau \in [0,\infty)\}.$$0. The strongest austral-summer SAD signals occur for initiation in phases 2–3, when dry SACZ and wet SESA appear already in phases 2–3 and intensify in phases 4–5. El Niño acts as a conditioning state: for 2–3 initiations it reinforces and prolongs the SAD, while for 6–7 initiations it enhances SACZ wet anomalies in phases 8–1. The study does not translate phase progression into fixed day lags and advises using the phase sequence as an ordinal lead–lag structure rather than a deterministic timing rule (Minjares et al., 27 Jul 2025).

The principal limitation across the literature is definitional nonuniformity. One study operationalizes SAD through subregional precipitation persistence in Amazonia, while the other diagnoses it qualitatively through OLR and upper-level circulation over the SACZ–SESA sector. This means that SAD is better treated as a family of dipolar rainfall responses than as a single metric. A related misconception is that any wet–dry contrast over South America automatically constitutes the same SAD; the papers instead show that the geometry, forcing pathways, and diagnostic variables are domain-specific.

Additional limitations are explicit. The Amazonian study does not report correlation coefficients, regression slopes, p-values, or confidence intervals, and its lead–lag relationships are qualitative. The SST anomaly baseline and climatological seasonal cycle are not specified in detail, and the paper does not present direct vertical-motion or moisture-flux diagnostics. The MJO study does not define a formal SAD index, does not provide mm day$$DPL(t) = \tau,\qquad \tau = \min\{\tau: A(t+\tau)\ge 0,\ \tau \in [0,\infty)\}.$$1 precipitation anomalies, and relies on OLR as a rainfall proxy together with 200 hPa fields. Its significance testing uses a 90% two-sample difference t-test, and later-phase or ENSO-stratified composites can be noisy because sample sizes become small.

The cumulative picture is therefore precise but bounded. The South American Rainfall Dipole is a recurrent, physically interpretable opposition of rainfall anomalies over South America, but its exact manifestation depends on the spatiotemporal window under study. In Amazonia it is expressed as Pacific-linked NA dryness versus Atlantic-linked SA dryness; in austral summer it is expressed as SACZ–SESA opposition whose strength and sign are modulated by MJO phase, event extremeness, and ENSO background state.