Skip Navigation Links weather.gov 
NOAA logo-Select to go to the NOAA homepage National Weather Service Forecast Office   Select to go to the NWS homepage
Pacific ENSO Applications Climate Center
banner piece
  banner piece
Local forecast by
"City, St" or Zip Code
Pacific ENSO Update

3rd Quarter, 2011 Vol. 17 No. 3


SEASONAL SEA LEVEL OUTLOOKS FOR THE U.S.-AFFILIATED PACIFIC ISLANDS


The following sections describe: (i) the Canonical Correlation Analysis (CCA)-based forecasts of sea-level deviations for forthcoming seasons JAS, ASO, and SON of 2011, (ii) the observed monthly mean and maximum sea-level deviations for the season AMJ 2011, (iii) a synopsis of sea level rise and enhanced trade wind with suggestions for further reading on the subject, (iv) forecast verifications for JFM 2011 (observed/forecast values), and (v) tide predictions. Note that the deviations are defined 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’. Also note that the CCA-forecasting technique adapted here does not account for sea-level deviations created by other atmospheric or geological factors such as tropical cyclones, storm surges or tsunamis.

tide station location
 

(i) Seasonal sea level forecast (deviations with respect to climatology) for JAS, ASO, and SON of 2011

Forecasts of the mean and maximum sea-level deviations in the USAPI are presented using CCA statistical model. Based on the independent SST values in AMJ 2011, the resulting CCA model has been used to forecast the sea level of three consecutive three-month periods: JAS, ASO, and SON of 2010 (see Table 1). CCA cross-validation forecast skills for 0, 1, and 2-month leads are presented in Fig. 3.

The forecast values of sea-level for JAS, ASO, and SON displays a positive deviation in the vicinity of north and south Pacific Islands, except one station in Majuro (for JAS). All the tide gauge stations (at 0 to 2-months lead time) show skillful forecasts for these three consecutive seasons (Table 1: bottom panel). The 2010 /11 La Nina condition has ended and ENSO-neutral condition now prevail. The atmospheric component of the previous La Nina was so strong and long-lasting that we still have some impact of La Nina. Consistent with this La Nina impact (i.e., enhanced trade wind), the sea level in these islands is still higher than normal.

Table 1 : Forecasts of MEAN and MAXIMUM sea level deviation in inches for forthcoming seasons

 
Seasonal Mean Deviations1
Seasonal Maximum Deviations2
Tide Gauge Station
JAS
ASO
SON

Forecast Quality3

JAS
ASO
SON
Forecast Quality3
RP for JAS Season4
Lead Time5
0
1M
2M
0
1M
2M
20 yr
100 yr
Marianas, Guam
+6
+4
+2
Good
+22
+20
+19
Good
6.3
10.9
Malakal, Palau
+3
+3
+2
Very Good
+40
+39
+38
Very Good
8.1
10.2
Yap, FSM
+3
+2
+2
Very Good
+30
+30
+29
Very Good
8.4
11.3
Chuuk, FSM**
+3
+2
+2
N/A
+30
+30
+30
N/A
N/A
N/A
Pohnpei, FSM
+1
+1
+1
Very Good
+29
+29
+30
Very Good
5.8
7.0
Kapingamarangi, FSM
+2
+2
+2
Good
+27
+27
+28
Fair
3.5
4.2
Majuro, RMI
0
+1
+2
Fair
+38
+40
+41
Fair
5.2
6.8
Kwajalein, RMI
+1
+1
+1
Fair
+38
+38
+38
Fair
4.1
5.2
Pago Pago, AS
+4
+4
+4
Very Good
+28
+28
+27
Good
4.1
5.4
Honolulu, Hawaii
+2
+2
+3
Fair
+21
+20
+21
Fair
3.4
5.7
Hilo, Hawaii
+2
+3
+2
Fair
+25
+24
+24
Fair
6.4
7.7

Remarks: As compared to the previous seasons, the forecasts values of sea level for JAS, ASO, and SON seasons (Table 1, above) indicate a negative trend (fall) in the months to come. However, currently, most of these stations show about 2-4 inches higher than normal sea level.

Despite an elevated sea level for about 2-4 inches higher than normal in the forthcoming seasons, the forecasts clearly indicate a fall in the same time horizon. This falling trend is supportive to on-going ENSO-neutral condition. According to CPC, a transition to ENSO-neutral conditions is expected to develop during May-June 2011 and continue through the Northern Hemisphere summer 2011.

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; also note that any deviations from -1 to +1 inches are considered negligible and deviations from -2 to +2 inches are unlikely to cause any adverse climatic impact. Forecasts for Chuuk (**) are estimated subjectively based on information from WSO Chuuk and observations from neighboring stations of Pohnpei and Yap.

1 Seasonal Mean Deviations is defined as the difference between the mean sea level for the given month and the 1975-1995 mean sea level value at each station. Likewise, 2 Seasonal Maximum Deviations is defined as the difference between the maximum sea level (calculated from hourly data) for the given month and the 1975-1995 mean sea level value at each station.

3 Forecast Quality is a measure of the expected CCA cross-validation correlation skill. In general terms, these forecasts are thought to be of useful (but poor) skill if the CCA cross-validation value lies between 0.3 ~ 0.4 (Fig. 3). Higher skills correspond to a greater expected accuracy of the forecasts. Skill levels greater than 0.4 and 0.6 are thought to be fair and good, respectively, while skill levels greater than 0.7 are thought to be very good.

4 Return Period (RP) of extreme values is calculated from hourly sea-level data. For example, the predicted rise of 6.3 inches at 20-year RP at Marianas, Guam indicates that this station may experience an extreme tide event once every 20 years that could result in sea-level rise of up to 6.3 inches above the median of seasonal maxima during the JAS season. Likewise, about once every 100 years we can expect the highest JAS tide at Marianas, Guam to be as high as 10.9 inches above the median of seasonal maxima. During some seasons some stations display alarmingly high values at the 20 and 100 year RP. These high values are due to large and significant increases in the tidal range caused by the passage of past storm events during that season. Click here to view probability of exceedence graphs for the JAS season.

5 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 (JAS), 1 (ASO), and 2 (SON) month leads based on SSTs of AMJ 2010.

 

(ii) Observed monthly sea level deviation in AMJ 2011 Season

The monthly time series (AMJ 2011) for sea-level deviations have been taken from the UH Sea Level Center. Note that ‘deviation’ is defined here as ‘the observed or forecast difference between the monthly mean [or maximum] and the climatological monthly mean values (from the period 1975- 1995) computed at each station’. Locations of all these stations are shown in Figure 2 (top of page).

 

Table 2: Monthly observed MEAN and MAX sea level deviations in inches for April, May, and June, with year to year standard deviations (SD).

Tide Gauge Station

(1) Monthly Mean Deviations
(2) Monthly Maximum Deviations
Apr
May
Jun
SD
Apr
May
Jun
SD
Marianas, Guam
*
*
*
3.7
*
*
*
3.9
Malakal, Palau
+5.4
+2.0
*
4.0
+40
+37
+32
3.8
Yap, FSM
+2.2
+1.3
+2.4
3.4
+31
+29
+26
4.0
Chuuk, FSM **
*
*
*
*
*
*
*
*
Pohnpei, FSM
+6.5
+5.1
*
2.2
+38
+34
*
2.7
Kapingamarangi, FSM
*
*
*
2.8
*
*
*
3.1
Majuro, RMI
+4.3
+4.5
*
1.8
+45
+42
*
2.9
Kwajalein, RMI
+3.2
+3.8
+3.9
2.3
+43
+38
+37
2.7
Pago Pago, AS
+10.3
*
*
3.7
+37
*
*
4.2
Honolulu, Hawaii
0.0
-1.0
0.0
1.8
+20
+20
+20
1.9
Hilo, Hawaii
+1.0
+0.5
+1.5
2.0
+21
+24
+24
2.4

* Data currently unavailable; 1 Difference between the mean sea level for the given month and the 1975 through 1995 mean sea level value at each station; 2 Same as 1 except for maxima; SD stands for standard deviations.

Remarks: As compared to May 2011, the monthly mean sea-level in June 2011 didn’t change much, except for Yap where it recorded slight rise. Currently, all stations are slightly higher than normal. The fall in Palau is quite considerable. There are several missing data; based on sea-level data on April, Pohnpei recorded fall of sea-level while Majuro recorded a marginal rise of sea-level. The monthly maxima also displayed a falling trend everywhere. Currently, the sea-level in the north Pacific is about 2-4 inches higher than normal. Most of the stations displayed about 2-4 inches fall of sea level during the last two months. This trend is supportive to on-going weakening stage of La Niña. At this stage our data seem to be very supportive of an ENSO-neutral condition.

 

(iii) Sea Level Rise and Enhanced Trade Wind: A Synopsis

While the sea level variations in the USAPI region are highly sensitive to the ENSO-cycle (i.e., low sea level during El Niño and high sea level during La Niña), the sea level rise during the La Niña year of 2007-08 was considerably higher (2 to 6 inches) than previous La Niña years at several locations. Despite somewhat smaller Oceanic Nino Index (ONI) values (indicating a weaker La Niña), as compared to ONI values during the 1998-99 and 1988-89 events (relatively stronger La Niña events), most of the islands recorded higher sea levels during the 2007-08 event (Chowdhury et al., 2010b). Based on this loosely proportional relationship between El Niño-Southern Oscillation (ENSO) strength and sea level anomaly, reasoning suggests that, in addition to the La Niña of 2007-08, there must have been other factors responsible for this rise.

One possible explanation is that the recent trend of enhanced trade winds west of the dateline is partly responsible for this rise (Timmermann et al., 2010; Merrifield, 2011 and references within). Whether these enhanced trade winds are abnormal or a longer-term trend remains to be seen. However, this finding is contradictory to the IPCC-Fourth Assessment Report (IPCC-AR4), which tended to show a weakening tropical circulation in response to an imposed warming signal. A major difference between these two types of results is one of time scales. The trend is sensitive to the start and end points of the time series analyzed. While the models used in IPCC-AR4 are on longer time scales, the others are on shorter time scales. In shorter times scales it is clear that enhanced trade winds west of the dateline are causing sea levels to rise up. In longer time scales, it is unclear to what extent the rise is a reflection of natural variability or a longer-term trend.

References and for Further Reading:
Chowdhury M. R., Barnston A. G., Guard C., Duncan S., Schroeder T, and Chu P-S 2010b. Sea level variability and change in the U.S-Affiliated         Pacific Islands - Understanding the high sea levels during 2006–08, Weather 65(10), 263-268.
Merrifield, M. A., 2011. A shift in western tropical sea-level during the 1990s, J. Clim. (forthcoming).
Timmermann, A., McGregor S., and Jin F-F. 2010. Wind effects on past and future regional sea level trends in the South Indo-Pacific, J Climate 23, 4429-4437.

 

(iv) Forecast Verification (Seasonal Mean) for AMJ 2011

Observed and forecast seasonal sea level values for the AMJ 2011 season are presented in Figure 4. Forecasts were in general skillful. However, Yap is over forecasted and Majuro is under forecasted. Some other stations are also not well forecasted. Missing data for these stations might have contributed this verification process. Added to this, stronger than expected atmospheric component (kind of La Nina type) in a ENSO-neutral season may be a cause for recent forecast anomaly.

 

(v) Tide Predictions (July 1 to September 30, 2011)

NOAA's web site for tide and currents has been used to generate the water level plot for the next four months. Predicted water level plots from July 1 to September 30, 2010 for five stations [(a) Marianas, Guam (b) Kwajalein, RMI (c) Pago Pago, American Samoa (d) Honolulu, Hawaii and (e) Hilo, Hawaii] are provided below.

 

Figure 5 (below): Predicted water level for the AMJ 2010 season at (a) Marianas, Guam (b) Kwajalein, RMI (c) Pago Pago, American Samoa, (d) Honolulu, HI and (e) Hilo, Hawaii. Data from NOAA/NOA/CO-OPS. X-axis: date/time (GMT); Y-axis: height in feet relative to Mean lower low water level (MLLW); MR: Mean-difference between high and low; SR: Difference between high and low tide during full moon (spring tide); and ML: Arithmetic means of high and low tides.

 

Predicted Water Level Plot for Guam

a) Marianas, Guam

 

Predicted Water Level Plot for RMI

b) Kwajalein, RMI

 

Predicted Water Level Plot for American Samoa

c) Pago-Pago, American Samoa

 

Predicted Water Level Plot for Honolulu

d) Honolulu, Hawaii

 

Predicted Water Level Plot for Hilo

e) Hilo, Hawaii



Pacific ENSO Applications Climate (PEAC) Center
c/o NOAA NWS - Weather Forecast Office Honolulu
2525 Correa Road, suite 250
Honolulu, HI 96822
(808) 956-2324

Web Master's email: peac@noaa.gov
Page Last Modified: August 06 2011 00:52:56 GMT

Disclaimer
Credits
Glossary

Privacy Policy
About Us
Career Opportunities