Pacific ENSO Update

4th Quarter, 2005 Vol. 11 No. 4

NEW - Experimental Sea level Forecasts
(deviations w. r. t. climatology) for the U.S-affiliated Pacific Islands

The following sections describe: (i) the CCA-based forecasts for sea level deviations for the forthcoming season, and (ii) the observed monthly sea level deviations. All units are in inches. Note that the forecasting technique adapted here does not account for sea level deviations created by other atmospheric or geological conditions such as tropical cyclones, storm surges or tsunamis.

Seasonal Sea Level Forecast for OND, NDJ, and DJF, 2005

Forecasts of the sea level anomalies in the USAPI are presented using CCA statistical model. Locations of all stations are shown in Fig. 3. Based on the independent SST values in JAS 2005, the resulting CCA model was used to forecast the sea level of three consecutive months: Oct-Nov-Dec (OND), Nov-Dec-Jan (NDJ), and Dec-Jan-Feb (DJF) (Table 1). CCA cross-validation forecast skills for 0, 1, and 2-month leads are presented (Fig. 4).


The OND season provided a more skillful forecast than NDJ and DJF. Most of the tide gauge stations show strong skill level ( Fig. 5). Only Guam, Saipan, and Pago-Pago of south Pacific display a relatively lower skill, which is still close to 0.5. All three consecutive periods (OND, NDJ, DJF) of the majority of the tide gauge stations are very well predicted with a mean skill greaterthan 0.60 (at 0 to 2-month lead time). In the previous issue of the newsletter, it was stated that the forecast skill for the season of JAS — which is based on spring SSTs in AMJ — is slightly less skillful because of the effect of a spring barrier. As the year advanced, the effect of the spring barrier gradually decreased and the forecasting skill increased in OND.

Table 1 : Forecasts of sea level deviation in inches (OND: Oct-Nov-Dec, NDJ: Nov-Dec-Jan, and DJF: Dec-Jan-Feb)

Tide Gauge Station OND NDJ DJF Forecast quality1

1. Forecast quality is a measure of the expected CCA cross-validation correlation skill. In general terms, these forecasts are thought to be of useful skill (or at least fair skill) if the CCA cross-validation value is greater than 0.3 (Fig. 4). Higher skills correspond to greater expected accuracy of the forecasts. Skill levels greater than 0.4 and 0.5 are thought to be moderate and good, while skill levels greater than 0.6 and 0.8 are thought to be strong and very strong respectively.

2. The lead time is the time interval between the end of the initial period and the beginning of the forecast period. For example, lead-0, lead-1M, and lead-2M means ‘sea-level’ of target season 0, 1, and 2 month leads based on SSTs of previous JAS.

Lead Time2 0 1M 2M  
Guam ** ** ** Good
Saipan ** -1 -1 Good
Malakal +1 ** -2 Very Strong
Yap ** -1 -1 Strong
Pohnpei +2 +1 +2 Very Strong
Kapingamari +3 +3 +3 Strong
Majuro +2 +2 +2 Strong
Kwajalein ** ** ** Strong
Pago Pago +2 +2 +2 Good

Note: (-) indicates negative deviations (fall of sea level from the mean), and (+) indicates positive deviations (rise of sea level from the mean), N/A: data not available. Deviations of +/-1 in. are considered negligible and denoted by **. Deviations +/- 2 in. are unlikely to cause any adverse impact.

Using a historical perspective, during OND 1997 and 1998, Yap recorded 8.6 inches of negative and 6.3 inches of positive deviations respectively. Pago-Pago recorded 1.5 inches of positive and 3.4 inches of negative deviations during the same period.

Observed monthly sea level deviation in JFM, 2005

The monthly time series (July to August) for sea level deviations have been taken from the UH Sea Level Center. The full time series (in mm) is available at ftp://ilikai.soest.hawaii.edu/islp/slpp.deviations. Deviations are defined here as the difference between the mean sea level for the given month and the 1975 through 1995 mean sea level value computed at each station.

Table 2 provides the monthly observed sea level deviations (in inches). All tide gauge stations recorded positive deviations in this quarter (3rd quarter) with Guam and Yap recording considerable positive deviations last quarter. As in previous quarters (JFM and AMJ), Guam continued to show sharp rise in this quarter too. The month of July and August displayed 12.3 and 10.5 inches of rise from the mean respectively (see Table 2 and Fig. 5; sea level data for the month of September is not available from the UH sea level center at print time). Malakal and Pohnpei recorded marginal to moderate positive deviations. Malakal initially remained steady in July, however, considerable variations were observed in August. Similarly, Kapingamarangi, Majuro and Kwajalein also provided positive deviations during the same time period. The sea level of the lone south Pacific station (American Samoa at Pago-Pago) displayed a rise during the same time period.

In last quarter’s issue, we forecasted positive deviations for all the tide gauge stations in JAS ( Fig. 5 dotted line). Real-time observed sea level data in JAS provided a very close match with the forecast values ( Fig. 5). All the observed values, except Guam and Yap, remained very close to the forecast ones. The deviations of Guam and Yap were considerably higher than the forecast values captured by CCA model simulation for prediction in JAS.
The relationship between the sea level and the SST time-series at each geographical grid-point has been investigated (figures are not reported here). Observations revealed, for example, the JAS sea level in Guam displays strong and positive correlations with the SSTs of the preceding and current season in the western Pacific. This positive correlation implies that warmer sea waters and more heat content with a deeper thermocline in the western North Pacific correspond to a higher sea level in Guam.

Table 2 : Monthly observed sea level deviations in inches (std deviation in parentheses)

Tide Gauge July August September3
  Deviation SD Deviation SD Deviation SD
Guam +9.4 (4.6) +10.9 (4.6) N/A (4.6)
Saipan N/A (2.9) N/A (2.4) N/A (3.1)
Malakal 0 (4.4) +4.1 (4.2) N/A (4.3)
Yap +9.2 (4.6) +9.7 (4.1) N/A (4.6)
Kwajalein +4.6 (2.2) +3.5 (2.0) N/A (2.7)
Majuro +1.1 (2.2) +1.1 (2.2) N/A (3.2)
Pohnpei +3.3 (2.7) +0.9 (2.7) N/A (3.7)
Kapingamar. +1.7 (2.3) +2.2 (2.5) N/A (2.6)
Pago Pago +1.7 (3.1) +3.7 (2.8) N/A (2.6)

Note: - indicate negative deviations (fall of sea-level from the mean), and + indicate positive deviations (rise of sea-level from the mean), n/a: data not available, Figures in parenthesis are year-to-year SD (standard deviations) for the month.

3. September data is unavailable at print time.