Michael W. Douglas1, Malaquias Peña2, Norma Ordinola3, Luis Flores3, Joshua Boustead4, and Jose Luis Santos5.

(1) NOAA/NSSL, Norman, OK USA

(2) CIMMS/Univ. of Oklahoma Norman, OK USA

(3) Universidad de Piura, Piura, Peru

(4) Creighton University, Omaha, NE USA

(5) ESPOL, Guayaquil, Ecuador



The rainfall over the coastal regions of northern Peru and Ecuador associated with strong El Niño events is one of the most dramatic climatic anomalies found anywhere on earth. Regions that are veritable deserts in most years, with little if any rainfall, can receive 1000 to 4000 mm of rainfall in less than 6 months. During 1997-8 an exceptionally strong El Niño event occurred, with anomalously warm water temperatures across the equatorial eastern Pacific and along the coast of Ecuador and Peru. Along the coast of northern Peru the rainfall was most intense during the months of January through April 1998.

Fig. 1 The pilot balloon network during Jan - May 1998

The evolution of the Sea surface temperature anomalies was well-monitored by the TOGA-TAO buoy network and satellite measurements, and it was possible to predict in general terms that the rainfall during the January-May 1998 period would be much greater than normal. The previous very strong El Niño event of 1982-3 was not as well anticipated, and the conditions responsible for the extremely heavy rainfall have only been described from surface and satellite measurements, as there were no meteorological soundings in the region. Because the PACS-SONET project had already established pilot balloon stations at Piura, Peru, and at Guayaquil, Esmeraldas, and San Cristobal (Galapagos Is), Ecuador in May 1997, there existed an opportuntity for somewhat enhanced monitoring of the phenomena in the region. In late 1997 it was recognized that the strength of the El Niño event would significantly bias the 6-month PACS-SONET observations that had been designed for other purposes, and and funds were requested to enhance the PACS-SONET network to take advantage of the unique opportuntiy to better describe the regional nature of the El Niño event. During December 1997 special pilot balloon stations were established at 2 additional locations in Ecuador and at 5 additional sites in Peru. In addition, more than 100 simple raingauges were installed throughout northern Peru and in western Ecuador to improve the description of daily rainfall over the region. The raingauge network consisted of volunteers, mostly individuals at Public Health Centers scattered throughout the countryside.


This study arose from the observation that there appeared to be, based on the rainfall observations, large fluctuations in the daily rainfall during the El Niño rainy period. The mean daily rainfall, determined from approximately 100 sites in northern Peru, showed a quasi-periodic fluctuation with approximately one week periodicity (Fig 2). The time series of frequency of stations reporting rain also showed a similar periodicity. Using this time series wet and dry days were identified. Then the mean winds at each pilot balloon station were calculated, as were the daily windfields from the NCEP reanalyses. Radiosonde observations made at the University of Piura have also been composited in a similar manner into wet and dry day means.

Fig 2. Daily mean rainfall (mm) for all stations in northern Peru during period Dec 1, 1997 - April 30, 1998. Smoothed curve suggests seasonal evolution.


Fig. 3 shows the difference between radiosonde observations made on wet and dry days at Piura, Peru during January 1998. The soundings are clearly more moist in the 2-8 km layer on wet days; whether this is a result of enhanced convection or is representative of conditions prior to the rainfall is not clear from this limited analysis.

Figure 4 shows the an average of the zonal and meridional wind profiles at five coastal pilot balloon stations in Ecuador and Peru during wet days and dry days. It is clear that during wet days there is relatively stronger westerly flow (towards the coast) up to 5 km above sea level. The southerly meridional flow on wet days is weaker below 2.5 km than during dry days, but is more southerly above this level, suggesting a meridional circulation cell. Examination of the wind differences for each station (not illustrated) showed large wet-dry

Fig. 3. Difference between wet and dry day radiosonde humidity profiles at Piura, Peru during January 1998.

Fig. 4 Average zonal and meridional wind profiles during wet and dry days at coastal pibal stations.

differences at San Cristobal in the Galapagos, indicating that the westerly wind anomaly during wet days extends far from the coast.

The NCEP reanalyses were examined to see whether they captured the windfield variations exhibited by the pilot balloon observations. The difference field at 850mb (Fig. 5) and other levels (not depicted) shows that during wet days the westerly wind anomalies are deep (from surface through 700mb level), and extend westward to at least 100ûW. There is reasonable agreement with the pilot balloon means along the coast and in the Galapagos Islands. It is encouraging that the NCEP reanalyses appear to describe the basic aspects of the synoptic-time scale variations in rainfall over northern Peru. This seems to indicate that satellite-based wind estimates and other sources of information (quite sparse conventional in-situ measurements) are adequate to describe the major circulation changes that are apparently associated with these synoptic time scale wet and dry variations.

Fig. 5. Difference at 850 mb between windfields during wet and dry days in northern Peru during January-April 1998. Maximum vector magnitude is 4.1 m/s.

Our collective results indicate that the influx of lower-tropospheric air from the Pacific Ocean is critical in producing the heavy rains during the El Niño period. This is in contrast to a relatively widely-held belief in the region that the moisture source is from the Amazon Basin, a notion that arises because the Infrared satellite imagery shows the cloudiness at mid and high levels to move from the east. In addition, the mean rainfall for the period January - April 1998, which is very similar to the mean pattern for the 1982-3 El Nino, shows that the rainfall anomaly with respect to climatology is largest west of the mountains and is relatively small east of the mountains. Such an anomaly pattern is difficult to reconcile with a moisture flux from the east.

Acknowledgments: The PACS-SONET program is supported by NOAA's Office of Global Programs. A more extensive acknowledgments can be found on the PACS homepage at: http://www.nssl.noaa.gov/projects/pacs

Corresponding author: Dr. Michael W. Douglas, NSSL, 1313 Halley Circle, Norman, OK, 73069 USA

email: mdouglas@nssl.noaa.gov