PosterPDF Available

Causes and Model Skill of the Persistent Intense Rainfall and Flooding in Paraguay during the Austral Summer 2015-2016

Authors:

Abstract

During the austral summer 2015-16, severe flooding displaced over 170000 people on the Paraguay River system in Paraguay, Argentina, and Southern Brazil. These floods were driven by repeated heavy rainfall events in the Lower Paraguay River Basin. Alternating sequences of enhanced moisture inflow from the South American Low-Level Jet and local convergence associated with baroclinic systems were conducive to mesoscale convective activity and enhanced precipitation. These circulation patterns were favored by cross-timescale interactions of a very strong El Niño event, an unusually persistent Madden-Julian Oscillation in phases four and five, and the presence of a dipolar SST anomaly in the central southern Atlantic Ocean. The simultaneous use of seasonal and sub-seasonal heavy rainfall predictions could have provided decision makers useful information about the start of these flooding events from at least two-to-four weeks in advance. Probabilistic seasonal forecasts available at the beginning of November successfully indicated heightened probability of heavy rainfall (90th percentile) over southern Paraguay and Brazil for December-February. Raw sub-seasonal forecasts of heavy rainfall exhibited limited skill at lead times beyond the first two predicted weeks, but a Model Output Statistics approach involving principal component regression substantially improved the spatial distribution of skill for Week 3 relative to other methods tested including extended logistic regressions. A continuous monitoring of climate drivers impacting rainfall in the region, and the use of bias-corrected heavy precipitation seasonal and sub-seasonal forecasts, may help improve flood preparedness for the austral summer season in this part of the world.
A31H-2289: Causes and Model Skill of the Persistent Intense Rainfall and Flooding in Paraguay during the Austral Summer 2015-2016
James Doss-Gollin1´
Angel G. Mu˜noz2Simon J. Mason3Max Past´en4
1Columbia Water Center, Columbia University 2AOS, Princeton University 3International Research Institute for Climate and Society, Columbia University 4Direcci´on de Meteorolog´ıa e Hidrolog´ıa, Paraguay
At a Glance
Beginning in Nov. 2015, repeated intense rainfall events associated
with mesoscale convective activity caused severe flooding along the
Paraguay-Paran´a river system (fig. 5, red box), displacing over
170000 people. We use a weather typing approach within a
diagnostic framework to show that:
IMoisture and energy advection via the South American Low-Level
Jet [SALLJ; 1], particularly during “No-Chaco” events [2] favored
mesoscale convective activity
IStrong El Ni˜no and active MJO favored a strong SALLJ but an
Atlantic dipole pattern influenced the jet’s exit region
INumerical forecasts predicted enhanced risk of heavy rainfall at the
seasonal scale, but biases in spatial patterns suggest difficulties
representing Pacific-Atlantic interaction
IUncorrected sub-seasonal model forecasts of rainfall had limited
skill beyond 10-15 days; use of Model Output Statistics –
particularly methods that correct both spatial patterns and
magnitudes – substantially improved forecast skill
Figure 1: Asuncion, Paraguay, 2015-16
Observations and Weather Types
Observations come from:
IRainfall: CPC Global Unified [3]
IAtmosphere: NCAR-NCEP Reanalysis II [4]
We use weather typing [5] to represent daily circulation patterns:
1. Calculate streamfunction Ψ from meridional and zonal wind [6]
2. Project 850 hPa streamfunction onto leading 4 EOFs
3. K-means clustering using classifiability index [7] to generate single
weather type for each day
Weather typing simplifies dynamics of daily rainfall but facilitates
analysis of sequences of daily weather patterns. They are
associated with patterns that have been well described in the
literature; particularly relevant are:
IWT1 represents “Chaco” jet event [8]
IWT4 represnts “No-Chaco” jet events [2]
Figure 2: Colors: anomalies of rainfall associated with each weather type
[mm d1]. Contours: anomalies of 850 hPa streamfunction [contour interval
1×104m2s1]
Observed Circulation Anomalies
Figure 3: Daily rainfall averaged over Lower Paraguay River Basin and observed
weather type for each day in NDJF 2015-16. Blue lines indicate the
climatological 50th, 90th, and 99th percentiles of NDJF area-averaged rain.
During austral summer (NDJF) 2015-16, most heavy rainfall
occurred during weather types 1 and 4 (fig. 3). Monthly-scale
circulation anomalies (fig. 4) show a weak anticyclonic circulation
that set up over central Brazil during November 2015 and
strengthened into the following month. In January 2016 it weakened
before returning in February 2016. The observed rainfall and
circulation anomalies are consistent with the aggregation of the
observed weather types shown in fig. 3.
Figure 4: Monthly composite anomalies observed during NDJF 2015-16. Colors:
anomalies of rainfall associated with each weather type [mm d1]. Contours:
anomalies of 850 hPa streamfunction [contour interval 1 ×104m2s1]
Lower Paraguay River Basin
Flat topography limits the river’s ability to carry the summer runoff,
causing seasonal inundation of the Pantanal and distributing the
river discharge in time [9, 10]. During NDJF 2015-16, river stage
[height] throughout the Lower Paraguay River Basin reached nearly
three times climatological levels (not shown).
Figure 5: Topographical map of the study area.
Acknowledgements
The authors thank David Farnham, Upmanu Lall, and Andrew
Robertson for insightful conversations and guidance. JDG thanks
the NSF GRFP program for support (grant DGE 16-44869). AGM
was supported by the Atmospheric and Oceanic Sciences (AOS)
Program at Princeton University.
S2S Model Forecasts
Figure 6: Chiclet diagram of ensemble-mean precipitation anomaly forecasts over
the Lower Paraguay River Basin from ECMWF S2S forecast data, as a function
of the forecast target date (horizontal axis) and lead time (vertical axis). Time
series of daily precipitation over the same area is plotted with y-axis inverted.
Figure 6 uses a Chiclet diagram [11] to visualize, as a function of
lead time, the time evolution of the uncorrected, ensemble-mean
rainfall anomaly forecast, spatially averaged over the Lower
Paraguay River Basin. At lead times greater than about two weeks,
the ensemble-mean forecast is slightly wetter than climatology. At
weather timescales (less than one week), the ensemble-mean
successfully predicts the timing and amplitude of the area-averaged
rainfall. At intermediate timescales, the model successfully forecast
the strongest breaks and pauses in the rainfall, such as the heavy
rainfall during December 2015 and the dry period during
mid-January 2016.
Model Output Statistics
We explore whether using Model Output Statistics [MOS; 12] can
improve the modeled representation of rainfall (fig. 7). Specifically,
we use: the raw model output (Raw); extended logistic regression
[XLR; 13]; heteroscedastic XLR [HXLR; 14]; principal component
regression [PCR; 15, 16]; and canonical correlation analysis [CCA;
15, 17] using 20 years of ECMWF forecasts.
Figure 7 indicates that better forecasts are obtained when both
magnitude and spatial corrections are performed (PCR and CCA).
The enhanced skill is achieved through the spatial corrections via
the EOF-based regressions, which – in contrast with the extended
logistic models – use information from multiple grid-boxes,.
Figure 7: MOS-adjusted S2S model forecasts and skill scores. Top row shows the
heavy rainfall (>90th percentile exceedance) forecast for 1-7 December 2015 as
the odds ratio oddsrp
(1p)
(1pc)
pc. Second row shows the Ignorance score
IGN log2p(Y). Bottom row shows the 2AFC skill score for each grid cell.
For all three rows, the grid cells which experienced a 90th percentile exceedance
for 1-7 December 2015 are outlined in black.
ENSO, MJO, and Weather Types
Figure 8: Anomalous probability of occurrence of each weather type concurrent
with observance of each MJO and ENSO phase. Only values which are
significant at p < 0.10 are shown.
During El Ni˜no years such as 2015-16, weather type 1 – related to
Chaco jet events – occurs more frequently for almost all MJO
phases, and particularly during phases 1, 2, and 5, consistent with
observations of the year under study. This agrees with previous
studies [i.e. 18] which find that the intensity and exact location of
the anomalous low-level anticylonic anomaly over central Brazil is
relevant for the precise impact of ENSO events.
Atlantic-Pacific Interaction and WT4
While the occurrence of WT1 during NDJF 2015-16 is well
explained by ENSO and MJO variability, these features alone do not
explain the occurrence of WT4, the “No-Chaco” jet event. Previous
studies emphasize the importance of Pacific-Atlantic interaction for
forecasting climate effects in this region [2, 5, 19, 20]. A persistent
SST dipole in the central southern Atlantic Ocean favors the
occurrence of WT4 by blocking transient extratropical wave activity
coming from the Pacific, facilitating transitions from Chaco jet
events (WT 1) to No-Chaco jet events (WT 4) via enhanced
low-level wind circulation from southern Brazil towards the Atlantic,
and back to north-east Brazil and the Amazon (see fig. 9) due to
land-sea temperature contrasts. Composite analysis (not shown) of
months with many WT4 occurrences is consistent with the
schematic shown here.
No or Weak Central South Atlantic SST dipole Central South Atlantic SST dipole
b
a
Eq Eq
30S
30S
Figure 9: Simple schematics of low-level jet events (red arrows) during austral
summer and El Ni˜no years.
References
[1] J. A. Marengo et al. International Journal of Climatology (2012). [2] C. Vera et al.
Bulletin of the American Meteorological Society (2006). [3] P. Xie et al. “CPC Unified Gauge-
Based Analysis of Global Daily Precipitation”. Preprints, 24th Conf. on Hydrology, Atlanta, GA,
Amer. Meteor. Soc. 2010. [4] M. Kanamitsu et al. Bulletin of the American Meteorological
Society (2002). [5] ´
A. G. Mu˜noz et al. Journal of Climate (2015). [6] A. Dawson. Journal of
Open Research Software (2016). [7] P.-A. Michelangeli et al. Journal of the Atmospheric Sciences
(1995). [8] P. Salio. Journal of Geophysical Research (2002). [9] J. M. Bravo et al. Journal
of Hydrologic Engineering (2011). [10] V. R. Barros et al. Journal of Hydrometeorology (2004).
[11] G. W. Carbin et al. Bulletin of the American Meteorological Society (2016). [12] H. R.
Glahn and D. A. Lowry. Journal of Applied Meteorology and Climatology (1972). [13] D. S.
Wilks. Meteorological Applications (2009). [14] J. W. Messner et al. Monthly Weather Review
(2014). [15] S. J. Mason and O. Baddour. Seasonal climate: forecasting and managing risk
(2008). [16] D. S. Wilks. Meteorological Applications (2006). [17] A. G. Barnston and C. F.
Ropelewski. Journal of Climate (1992). [18] I. Velasco and J. M. Fritsch. Journal of Geophysical
Research: Atmospheres (1987). [19] M. Barreiro. International Journal of Climatology (2017).
[20] C. Lima et al. Earth System Dynamics Discussions (2017).
Corresponding author: james.doss-gollin@columbia.edu
ResearchGate has not been able to resolve any citations for this publication.
  • J A Marengo
J. A. Marengo et al. International Journal of Climatology (2012). [2]
  • C Vera
C. Vera et al. Bulletin of the American Meteorological Society (2006).
CPC Unified Gauge-Based Analysis of Global Daily Precipitation
  • P Xie
P. Xie et al. "CPC Unified Gauge-Based Analysis of Global Daily Precipitation". Preprints, 24th Conf. on Hydrology, Atlanta, GA, Amer. Meteor. Soc. 2010. [4]
  • M Kanamitsu
M. Kanamitsu et al. Bulletin of the American Meteorological Society (2002).
  • P Salio
P. Salio. Journal of Geophysical Research (2002). [9]
  • J M Bravo
J. M. Bravo et al. Journal of Hydrologic Engineering (2011).
  • V R Barros
V. R. Barros et al. Journal of Hydrometeorology (2004).
  • G W Carbin
G. W. Carbin et al. Bulletin of the American Meteorological Society (2016). [12] H. R. Glahn and D. A. Lowry. Journal of Applied Meteorology and Climatology (1972). [13] D. S.
  • J W Messner
J. W. Messner et al. Monthly Weather Review (2014). [15] S. J. Mason and O. Baddour. Seasonal climate: forecasting and managing risk (2008). [16]
Meteorological Applications (2006)
  • D S Wilks
D. S. Wilks. Meteorological Applications (2006). [17]