Multi-parametric analysis of an El Nino event using satellite derived SST, SSH and wind field measurements

Abstract
The E1 Niflo event was investigated during NOAA AVHRR Multi-Channel Sea-Surface Temperature (MCSST) data, ERS AMI wind scatterometer data and TOPEX/Poseidon Altimeter data in the Pacific Ocean from 1991 to 1997. Derived parameters (sea surface temperature, wind vector and Sea surface height) were analyzed in order to investigate the relation between the ocean and the atmosphere.

Measurements from al instruments are observed and show marked variations between E1 Nino and Non E1 Nino conditions. Further investigations of temporal variations between their main parameters suggest that earliest detection of an E1 Nino is provided by scatterometer.

The sea surface temperature, wind, sea surface height differs form E1 Nino and Non-E1 Nino season. SST anomaly shows that during the El Nino season, the temperature (approximately +5°C) is higher than the normal season. Observed wind pattern shows a difference between the peak of an E1 Nino year and normal year. The weakening of trade wind is apparent from the zonal wind analysis. Sea surface height form TOPEX/Poseidon altimeter data shows that the sea surface height (almost 20 cm) is higher than normal season in Eastern Pacific ocean during the E1 Nino season.

Introduction
In the recent study by NOAA and FAO, it was reported that the current E1 Nino event is very strong and the impact on the Asian countries are increasing and linked to the some known disasters. So, some kind of prediction of El Nino is necessary to lessen the range of the disaster.

Now, we study this event using many methods such as ship observing data, buoy data and satellite data. However, the use of accurate information is vital because the prediction and assessment of the El Nino, because many numerical prediction models are needed accurate information (Moore 1995). Particularly, the condition of El Nino cannot be understood using only one climate parameter. Hence, investigation should consider other factors such as sea surface height. If these measurements are studies within same space and time, the results can be improved to understand and predict E1 Nino.

Methodology
Sea Surface Temperature (SST) was taken from NOAA AVHRR Multi-Channel Sea-Surface Temperature (MCSST) data. The AVHRR data was provided as a weekly data set. But some data sets had large sparse area because of cloud, which created a difficulty in investigating the data. This problem was solved by merge the weekly data sets as a monthly data set.

SST anomaly was calculated, which is useful technique when the date are subject to seasonal variations and EOF (Empirical Orthogonal Function, Kutzbach 1967) was conducted on anomaly data.

ERS 1, 2 AMI wind scatterometer data was taken for Wind field data. The first alias wind speed data was used from the wind scatterometer data. Each data put on 25 km x 25 km geographical data box and a monthly mean wind speed, zonal wind and wind vector data set were created.

Sea surface Height data was extracted from the TOPEX/Poseidon Altimeter data. From the extracted altimeter data set, sea surface height data was calculated by subtracting the corrected range that affect altitude. The environmentally corrected TOPEX/POSEIDON sea surface height relative to the reference ellipsoid is defined by:

SSH= sat_height – (altimeter_range + delta_alt_range + altimeter_bias) – geoid_height -geocentric_body_tides-ocean_tides-atmospheric_loading

CNES (Centre national d'Etudes Spatiales) orbit altitude data was used as a satellite height. Delta_alt_range is the atmospheric corrections (wet troposphere correction, dry troposphere correction, ironospher correction, electromagnetic bias). The altimeter_bias is a bias of Topex/Poseidon altimeter. The TOPEX and POSEIDON ranges had been corrected for "absolute" bias based on platform measurements of sea level at Lampedusa Island and Harvest Platform. The best estimates of the respective absolute biases use the new orbits. In the data record, H_Geo was used as a geoid height. Geocentric_body_tieds is for body tides that are the movement of the Earth's crust which will cause changes in SSH. Ocean tides. H_EOT_C was used as a ocean tide. Atmospheric loading is the inverted barometer effect where an increase in atmospheric pressure results in a decrease in SSH, and an decrease in atm pressure results in an increase in SSH. After extracting all data set, the mean value of 95,96 and 97 data was used as a reference sea surface height. Fig 1 shows the study area and the TOPEX/Poseidon repeat cycle over study area.

Finally time series analysis and statistical analysis was conducted on 3 data sets.

Figure 1. TOPEX/Poseidon 10day repeat cycle tracks over the study area
Results and Discussion
Fig 2,3,4 shows the color-coded maps for the visual interpretation between the El Nino and non El Nino season.

Fig 2. Observed Monthly SST anomaly (a) a normal year, (b) an El Nino year

Fig 3. Surface Zonal wind, (a) a normal year, (b) an El Nino year
Fig 2 shows the abnormal patterns of temperature between the normal and El Nino season. During the El nino year, the cold are, in front of South America, almost disappeared. And the warm water occupying the East Pacific Ocean.

From the wind vector data, only U (zonal) component data was drawn as a color map (Fig 3). The weakening of easterly trade winds is evident between the El nino season and non El nino season.

Fig 4. Sea surface height, (a) a normal year, (b) an El Nino Year

Fig. 5. The surface Height behavior during the 2 years.
Fig 4 shows the sea surface height. During the El Nino season, in front of Eastern Pacific Ocean sea surface height increased significantly around 20 cm higher than normal season.

Tropical pacific ocean is divided 5 different regions El Nino 1, 2,3,4 and Warm pool area. These area was studied respectively except warm pool area which has much noise. Fig 5 shows the sea surface height change among the Nino 1-4 region. As you can see, the SSH on Nino 1, 2, 3 region was increased but, Nino 4 region was decreased during he El Nino season.

Fig. 6 Time series evolution of (a) SST, (b) zonal wind, (c) SSH

Fig. 7. Scatter plot and regression line for (a) SST-Wind, (b) SST-SSH, (c) Wind-SSH
Fig 6. Shows the time series evolution of Sea surface temperature (SST) anomaly, zonal wind, and sea surface height (SSH) on the Nino 3 region. The SST and SSH was relatively increased but the zonal wind decreased during the El Nino season.

Fig. 7 shows the regression analysis results. The sea surface height is positively related to sea surface temperature and negatively related to zonal wind.

Conclusion
Potential of satellite data in assassin the El Nino event were studied and the relationship among the sea surface temperature, wind and sea surface height were investigated using visual interpretation and statistical methods.

During the El Nino season, the sea surface temperature of Eastern Pacific Ocean was increased and the sea surface height was increased significantly. The zonal wind over whole pacific ocean was disrupted, the trade wind direction was changed and the western pacific ocean wind speed was weakened.

The regression analysis shows that the sea surface height is positively related to sea surface temperature and negatively related to zonal wind.

Acknowledgements
The author would like to express thanks for the data providing and information for this research:

• European Space Agency (ESA)
• Centre ERS dArchivage et de Traitment (CERSAT)
• Physical Oceanography Distributed Active Archive Center (PO. DAAC)
• Jet Propulsion Laboratory (JPL)
• Research Systems Incorporation (RS Inc)

Reference

• The future of Spaceborne Altimetry Ocean and Climate Change; A report prepared by the Future Altimetry Working Group
• W. CULDIP, 1994, Overview of altimater data processing at the U. K. Earth Observation Data Centre, International Journal of Remote Sensing, VOL. 15, NO. 4, pp. 871-887.
• McClam, E. P., W. G. Pichel, and C. C. Walton, 1985. Comparative Performance of AVHRR-Based Multichannel Sea Surface Temperatures, journal of Geophysical Research, 90. pp. 11587-11601.
• An Atlas of monthly Mean Distributions of SSMI Surface Wind Speed, AVHRR Sea Surface Temperature, AMI Surface Wind Velocity, and TOPEX/POSEIDON Sea Surface Height During 1995, JPL Publication 98-5.