Cover. Monitoring water surface ratio in the Chinese floods of summer 1998 by DMSP-SSM/I

The Strait of Bab al Mandab connects the Red Sea with the Gulf of Aden and Indian Ocean. Flow through the Strait is dominated by monsoonal forcing with surface water inflow in the winter and outflow in the summer. Analysis of TOPEX/POSEIDON altimeter sea surface height residuals at subsatellite locations in the Red Sea, the Strait and Gulf of Aden indicates a dominant annual cycle and a secondary semi-annual cycle at all locations. A combined annual and semi-annual model is then fitted to sea surface height anomalies at each of the locations. The combined model explains most of the variance in the data: from 74 per cent variance in the Gulf of Aden to 92 per cent variance in the southern Red Sea. The amplitude of the annual cycle is 18cm in the Red Sea and 13 cm in the Gulf of Aden. An analysis of coherence between altimeterderived wind stress and sea surface height residuals shows that the annual cycle is probably related to wind forcing. A weaker semi-annual cycle of 4-8 cm is probably related to the cycle of evaporation. A two-layer model of flow through the Strait shows that the dynamic sea surface height signature is only 2cm. This signature of exchange is too small to be resolved by current satellite altimeter measurements     

Ocean-atmosphere coupling in the Mediterranean Sea from TOPEX/POSEIDON,ERS1 and AVHRR data

The objective of this work is to evaluate the patterns of the oceanatmosphere coupling in the Mediterranean sea using the statistical analysis of the time series of three satellite measured variables: wind stress curl computed from the surface wind velocities measured by the European Remote Sensing SatelliteERS1, sea levelanomalies from the merged ERS1-TOPEX/POSEIDON altimetric data and the sea surface temperature from the NOAA/NASA Pathfinder AVHRR Oceans Project. This study examines the ocean-atmosphere coupling patterns in time and space through a canonical correlation analysis of the fields. The wide and exhaustive coverage of the oceanic and atmospheric fields from satellites (such as TOPEX/POSEIDON and ERS1) allow a potential detailed inspection of the coupling. In this study we assess the role of wind stress and/or sea surface temperature space/time variations as forcing mechanisms of the sea level variability for the total 440 day duration (from October 1992 to December 1993) of the three satellite data sets in the different Mediterranean sub-basins. The linear barotropic vorticity equation is computed in order to estimate if the Mediterranean sea level response to wind forcing is barotropic. Results indicate that except for the Strait of Sicily and the Adriatic sea the barotropic response is not dominant. From the canonical correlation analysis, a correlation of 0.57 for the first mode between the sea level anomaly and the wind stress curl indicates a significant local coupling between them, especially in areas of the Eastern Mediterranean. Two uncoupled annual cycles are found in the sea surface temperature data.     

Sea level and geostrophic current features from tandem TOPEX/Poseidon‐Jason data in the Newfoundland offshore

We have investigated sea level and surface currents features over the Newfoundland Shelf and Slope using the tandem TOPEX/Poseidon (T/P) and Jason altimetry data (2002–2003). The consistency and error characteristics of T/P and Jason measurements are examined not only in terms of sea level and cross‐track current anomalies but also with respect to current anomalies at crossovers and the Labrador Current transport. Nominal absolute currents are constructed by adding the altimetric geostrophic current anomalies to a climatological‐mean model circulation field. The comparison of the sea level and cross‐track current anomalies from January to July 2002 shows overall good agreement between T/P and Jason, with correlation coefficients different from zero at the 95% confidence level at almost all locations for sea level and at most locations for currents. Errors are estimated to be 2.5 cm for sea level and 10 cm/s for cross‐track current anomalies. Analyses of the current variability at crossovers indicate approximate agreement of T/P and Jason measurements, except for the Northeastern Newfoundland Shelf and Slope probably due to the ice presence during the period. Model‐altimetry combined absolute currents are used to estimate near‐surface transport associated with the shelf‐edge Labrador Current, showing good correlation between T/P and Jason estimates and strong seasonal changes. The cross‐track geostrophic current anomalies from September 2002 to December 2003 are used to calculate the root‐mean‐square (rms) current variability at crossovers and to derive the shelf‐edge Labrador Current. The interleaved T/P and Jason observations can better capture the spatial distribution of shelf and slope circulation variability.     

Investigation of sea level variability in the Baltic Sea from tide gauge,satellite altimeter data,and model reanalysis

The exploration of sea level variability from multiple satellite altimetry measurements in the Baltic Sea is yet to be performed due to the lack of reliable data, especially near the coast. In the present study, the number of available altimeter measurements was increased by 96% in coastal regions by replacing the radiometer wet troposphere correction with model-based correction. The satellite altimeter data show good agreement with monthly tide gauge data, with mean a root mean square difference (RMSD) of 3.4 cm and a mean correlation coefficient of 0.97 at 42 tide gauges in the Baltic Sea, and higher RMSD and lower correlation coefficient in the Gulf of Bothnian and the Danish Straits, respectively. Independent model reanalysis demonstrates similar performance at the sites. The collocated high-frequency tide gauge data and altimeter data demonstrate an RMSD of 8.86 cm and a bias of 6.63 cm at Stockholm, respectively. The tide gauge sea level variations in the Baltic Sea are highly correlated with the Hurrell North Atlantic oscillation (NAO) index at the decadal timescale. A decrease in the correlation coefficient of the tide gauge winter averaged sea level time series and the winter Hurrell NAO index has been observed on the decadal timescale in the last decade, especially during 2003 and 2005.     

Impacts of altimeter corrections on local linear sea level trends around Taiwan

Multi-mission altimeter measurements from TOPEX, Jason-1, and Jason-2 satellite altimetry over the 1993–2009 time span are used to characterize the local linear sea level trend (LSLT) around Taiwan. The results show that the long-term changes of default geophysical and range corrections, i.e. the inverted barometer correction, wet tropospheric correction, and sea state bias correction, have significant impacts on the determination of local LSLT. The trend of default corrections contribute more than 1.4 mm year^?1 along the coastline of China mainland and 2.1 mm year^?1 to local LSLT in the Taiwan Strait. The default-corrected altimetric data exhibit highest and lowest local LSLTs in the southeast and northwest of Taiwan, respectively. The regional LSLTs of 3.8 mm year^?1 and 4.6 mm year^?1 are estimated from the default-corrected and uncorrected altimetric data in the study area, respectively.     

Operational use of ERS-1 SAR images in the Canadian ice monitoring programme

ERS-1 SAR images have been used to monitor sea ice in the Canadian east coast and the Gulf of St. Lawrence since December 1992. Operational ice parameters were obtained, in near real-time, through visual image analysis. The information was incorporated in standard daily operational products. Further analysis was conducted later to evaluate the data utility. The images are remarkably useful in discriminating between new ice and water and in delineating ice edges. Structural information of ice surface is usually smeared. Distinction between young ice and first-year ice types, based on backscatter alone, is not reliable. Rules for image interpretation were developed. They are mainly based on structural information in images, ancillary data on weather, and recent and historical ice records. Limitations on data utilization are explained in relation to the sensor and orbit parameters. An insight into operational requirements and their relation to SAR observations is presented.     

Technical note Empirical algorithm using SeaWiFS hyperspectral bands: A simple test

The aim of this note is to test an empirical algorithm using spectral curvature theory at selected SeaWiFS hyerspectral bands, in order to estimate the chlorophyll pigment concentrations in the coastal waters of Baie des Chaleurs (Gulf of St. Lawrence). Simulated SeaWiFS algorithms require the measurement of reflectances with a high sensitive spectroradiometer as well as the responsitivity of SeaWiFS sensor at each band. Volume reflectances were taken using a hand held spectroradiometer simultaneously to measurements of in situ chlorophyll pigment concentrations. An empirical algorithm of the form: log (chlorophyll pigments) 0.92-0.46 Delta2 log R0sea(555) yields an estimate of chlorophyll pigment concentrations within a mean deltaC / C of 23%.     

Observations of oceanic mesoscale features and variability in the Canary Islands area from ERS-1 altimeter data,satellite infrared imagery and hydrographic measurements

The main purposes of this paper are to identify and evaluate the oceanic mesoscale features that appear in the Atlantic Ocean eastern boundary, 20-40° N, 19-9° W, using ERS-1 (1992-1993) satellite altimeter data. The sea surface height anomalies and the eddy kinetic energy fields are calculated. High energy values, between 0.03 and 0.05 m 2 s -2, are observed with the altimeter data in the Canary region mainly in summer. These maximum values are associated with eddies located downstream of Gran Canaria and Tenerife (Canary Islands). Comparison with in situ measurements provided by the hydrographic surveys from a cruise in summer 1993 around the Canary Islands showed a good agreement. Dynamic heights relative to 300 dbar for August and the altimetric heights from ERS-1 data were averaged over 0.125° boxes for the duration of the cruise. The correlation coefficient was 0.7. Negative anomalies of the sea level calculated from ERS-1 between Cape Yubi and Cape Bojador (26.0°-27.5° N) were obtained in August 1993. Possibly these correspond with filaments from the north-west African upwelling coast. Also, the results of the altimetric data were compared with NOAA AVHRR (Advanced Very High Resolution Radiometer) sea surface temperature (SST) satellite images. A cyclonic and an anticyclonic eddy to the south-west of the island of Gran Canaria were identified during the same period in the SST images.     

Sea Ice Extent and Classification Mapping with the Moderate Resolution Imaging Spectroradiometer Airborne Simulator

An algorithm for mapping sea ice extent and generalized classification of sea ice by reflective and temperature characteristics with Moderate Resolution Imaging Spectroradiometer (MODIS) Airborne Simulator (MAS) data is presented. The algorithm was tested using a MAS scene over the Bering Sea near St. Lawrence Island, Alaska, USA, acquired 8 April 1995. Clouds were masked with the University of Wisconsin cloud masking algorithm. Ice surface temperature was estimated with a split-window technique. Sea ice extent and generalized type of sea ice were identified based on reflective characteristics and estimated ice surface temperature using a grouped criteria technique. Resulting maps were consistent with visual interpretation and with sea ice extent and type information reported in prior studies of the region.     

Seasonal-to-decadal modes of global sea level variability derived from merged altimeter data

Motivated by the first availability of more than 16 years (1992-2009) of merged altimeter data from up to four concurrent satellite missions with TOPEX/Poseidon, ERS-1 and 2, ENVISAT, Jason-1, and Geosat Follow-On, the global sea level is reexamined with special emphasis on its distinctive scales of variability. The global sea level variability is dominated by eleven statistically significant and geographically robust principal modes at seasonal-to-decadal timescales. The identified modes can be divided into four regimes, i.e., the seasonal regime with two principal components at 3 and 6 months, the annual regime with a single 12-month component of global dominance, the interannual regime with seven principal components at 1.55, 1.74, 1.94, 2.34, 3.07, 4.20, and 5.40 years, as well as the decadal regime with a 9.28-year subdecadal component. The frequency dispersion of the identified modes is found to be at ∼ 0% level for the annual regime, at ∼ 5% level for the seasonal and interannual regimes, and at ∼ 10% level for the decadal regime. In the space domain, the locations of principal modes are found to be geographically correlated. The seasonal regime is characterized by a small number of localized amplitude highs in regional seas. The annual regime is associated with major western boundary currents of the Gulf Stream and Kuroshio. For the interannual regime, mode active zones are observed in the western Pacific warm pool (WPWP), El Niño, and western Indian Ocean regions, among which the WPWP has a well-defined two-core structure, and the northern core coincides with the most energetic zone of the decadal regime. Another interesting finding is that east-west (zonal) and equatorial-extratropical (meridional) teleconnections are evident for the major signals in a rather symmetric and systematic pattern. Joint spatiotemporal analysis reveals that multi-modality at a given location and multi-regionality for a given mode are fundamental features in global sea level variability.     

Prediction of currents and sea surface elevation in the Gulf of California from tidal to seasonal scales

A web tool that provides currents and/or sea surface elevation in the Gulf of California is presented. The above variables are reconstructed from harmonic constants obtained from harmonic analyses of time series produced by a 3D baroclinic numerical model of the Gulf. The numerical model was forced (1) at the Gulf's mouth by the tides and the hydrographic variability of the Pacific Ocean (at semiannual and annual frequencies), and (2) at the Gulf's surface by winds, heat and fresh water fluxes (also at the semiannual and annual frequencies). The response to these forcings results in motions with time scales limited to semidiurnal and diurnal, fortnightly and monthly (due to nonlinear interactions of the tidal components), and semiannual and annual frequencies (due to the nontidal forcing).     

Satellite-derived conditions and advection patterns off Iberia and NW Africa: Potential implications to sardine recruitment dynamics and population structuring

Surface velocity estimates from wind and altimeter data, together with satellite-derived sea surface temperature and chlorophyll, were used to explore the advection patterns and environmental conditions using a simple Lagrangian model. Although the model is generic (only considering the physical transport), the results are of particular interest for sardine (Sardina pilchardus) early life stages, due to the spawning dynamics of the species off Iberia and NW Africa. Particles were released on the shelf of the Iberian and northern African Atlantic coasts every five days from 1998 to 2004 and advected for 27-days. Trajectories were computed using combinations of three surface velocity constituents: Ekman velocity, geostrophic velocity derived from sea level anomaly maps and mean geostrophic velocity derived from two mean dynamic topographies. Daily time series of sea surface temperature, chlorophyll-a concentration and water depth where constructed for each particle trajectory. The results showed strong regional and seasonal dependence of the transport due to wind, with average seasonal cycles of the percentage of particles found in the shelf reflecting the anticipated effect of the seasonal north–south migration of the trade wind belt. The addition of the geostrophic transport led to the attenuation of seasonal cycles and higher mean/maximum values in the probability of retention within the shelf. This increased capacity for retention even during strong upwelling conditions seems to have been overlooked by theories aiming to describe the reproductive strategies of pelagic fish and understand recruitment dynamics based primarily on wind variability. On the other hand, the average surface chlorophyll values indicated a distribution area with higher primary production for particles that never leave the shelf, in comparison to those that return to it or are found in the open sea a month after release, which is in agreement with existing hypotheses that offshore advection is detrimental to larval survival mainly due to the higher risk of starvation. Finally, the exchange between adjacent shelf areas was generally small (less than 5%) within the study period, with the exception of the Gulf of Cadiz. In the latter area, up to 50% of particles released in the southern Iberian shelf reached the Moroccan shelf in several events, contradicting previous suggestions that the Strait of Gibraltar acts as a physical barrier that promotes genetic differentiation among neighbouring sardine populations in the Atlantic. Overall, these results provide useful insights for the study of sardine dynamics in the northeast Atlantic but the unexpectedly high sensitivity of summary metrics (namely retention probability) to the choice of surface velocity field preclude firm conclusions and indicate alternative routes for future studies.     

The satellite-measured sea surface temperature change in the Gulf of Finland

Climate change in Baltic region and in the Gulf of Finland is an accomplished fact in human brains and in science. The purpose of this research is to retrieve quantitative level of changes for sea surface temperature (SST) of the Gulf of Finland. Two space systems National Oceanic and Atmospheric Administration/Advanced Very High Resolution Radiometer (NOAA/AVHRR) and Aqua/Moderate Resolution Imaging Spectroradiometer (MODIS) provided satellite data about temperature of the sea surface. SST data covers period 1981–2014 and includes 444 monthly data scenes with spatial resolution about 10 km. Data quality analysis displays high reliability of NOAA/AVHRR and Aqua/MODIS satellite information. The Gulf of Finland’s average annual SST has changed from 6.8°C in 1982 up to 8.2°C in 2014. Its mean speed of warming is about 0.04°C year^–1. The growth of the temperature was irregular, in the middle of 80th year, the temperature dropped down to 5.0°C, and then sharply increased up to 7.3°C in 1989. SST growth in the Gulf of Finland coincides with air temperature and sea temperature growth. The climate change in the Gulf of Finland has special significance due to the fragility of the northern ecosystems and high anthropogenic load.     

Seasonal and interannual variability of chlorophyll-a in the Gulf of Oman compared to the open Arabian Sea regions

Field sampling, remote sensing and modelling were employed to understand the seasonal and interannual changes of chlorophyll-a concentrations in the Gulf of Oman in comparison to open sea regions. In these regions, maximal chlorophyll concentrations were reported during the summer monsoon (with peaks in June and August), while in the Gulf of Oman, the chlorophyll maximum was observed during the winter monsoon (February–March). From 1997 through to 2008, the interannual variability in chlorophyll-a concentrations in the Gulf of Oman has not exhibited pronounced trends and neither have the other two (oceanic) regions in the western Arabian Sea. However, an increase of the annual variation in chlorophyll concentrations over the years was noticed. The diatom biomass decreased two-fold from 1997 to 2007. Nitrate concentration and mixed-layer depth also declined. In comparison to the seasonal blooms driven in the Gulf of Oman by the dinoflagellate Noctiluca scintillans, the year 2008 was markedly different. The summer bloom was shifted to September; it was gradually extended in time and formed by the other species. An applicability of the concept of ecosystem regime shift is discussed.     

Semi-annual cycle of sea-surface temperature in the East/Japan Sea and cooling process

Semi-annual cycles of the sea-surface temperature (SST) in the East/Japan Sea (EJS) and their forcings were examined by analysing National Oceanic and Atmospheric Administration/Advanced Very-High-Resolution Radiometer data, scatterometer wind vectors, and heat flux data. The semi-annual cycle contributed to the total variance of the SST by 8% and amounted to 25% of the amplitudes of the annual SST cycle, particularly in the Tatarskiy Strait and along the continental shelf off Russia. The lowest phase, corresponding to the minimum SST, occurred during early November and 6 months earlier in May or June depending on the position. The forcings of the semi-annual cycle were not semi-annual but substantially annual with a lag of 6 months, which gave rise to SST cooling in spring and autumn. Our analyses illustrated that SST cooling in autumn was caused by direct, local atmospheric wind forcings, whereas the cooling with large amplitudes of the semi-annual cycle in spring was caused by the non-local, remotely forced cold water advection of the Liman Current associated with sea-ice melting in the Tatarskiy Strait. The phase lag of 1–2 months between the complete melting of the sea ice in the Tatarskiy Strait and the surface cooling along the Russian continental shelf was related to the advection of cold water from sea ice in the form of the Liman Cold Current. The semi-annual cycle also resulted from asymmetry of the time series of the SST due to a long cold winter and a short warm summer. To understand how SST curves are distorted and asymmetric, we suggested two mathematical parameters of kurtosis and skewness. In addition, we suggest that the ratio of the semi-annual amplitude to the annual amplitude of SST harmonics can be used as a typical indicator of the asymmetry in year-to-year SST variations.     

The importance of coastal altimetry retracking and detiding: a case study around the Great Barrier Reef,Australia

A new approach for improving the accuracy of altimetry-derived sea level anomalies (SLAs) near the coast is presented. Estimation of SLAs is optimized using optimal waveform retracking through a fuzzy multiple retracking system and the most appropriate detiding method. With the retracking system, fuzzy-retracked SLAs become available within 5 km of the coast; meanwhile it becomes more important to use pointwise tide modelling rather than state-of-the-art global tidal models, as the latter leave residual ocean tide signals in retracked SLAs. These improvements are demonstrated for Jason-2 waveforms in the area of the Great Barrier Reef, Australia. Comparing the retrieved SLAs with in situ tide gauge data from Townsville and Bundaberg stations showed that the SLAs from this study generally outperform those from conventional methods, demonstrating that adequate waveform retracking and detiding are equally important in bringing altimetry SLAs closer to the coast.     

Seasonal and interannual variability of oceanographic processes in the Gulf of Guinea: An investigation using AVHRR sea surface temperature data

The Gulf of Guinea is situated in a critical position for understanding Atlantic equatorial dynamics. This study investigates seasonal and interannual variability in sea surface temperature (SST) throughout this region, focusing on dynamical ocean processes. A 10.5-year time series of remotely sensed SST data with 4 km spatial resolution from the Advanced Very High Resolution Radiometer (AVHRR) were used for this investigation, as they are sufficient to resolve shelf processes. Firstly, patterns of cloud cover were assessed, then spatio-temporal variability in SST patterns was investigated. Features identified in climatological SST images were the Senegalese upwelling influence, coastal upwelling, tropical surface water, river run-off and fronts. Of particular interest is a shelf-edge cooling along the coast of Liberia and Sierra Leone in February. Interannual variability, assessed using annual mean images, time series decomposition and spectral analysis, showed a quasi-cyclic pattern of warm and cool years, perhaps related to El Niño-type forcing. The results of this study show the usefulness of infrared remote sensing for tropical oceanography, despite high levels of cloud cover and atmospheric water vapour contamination, and they provide evidence for theories of westward movement of the upwelling against the Guinea current and remote forcing of the upwelling.     

Global land vegetation and marine fishery responses to atmospheric and oceanic decadal variability

The decadal variability of sea surface temperature (SST) and sea level pressure (SLP) anomalies, as well as the response of global land vegetation and marine fisheries, are investigated for three periods: 1982–1988, 1989–1998, and 1999–2008, separated by the 1988–89 and 1998–99 regime shifts. The goal is to develop a global-scale ecosystem concept to support an improved understanding of the corresponding changes in atmospheric, oceanic, and biological responses. The analysis is based on global SST, SLP, precipitable water content (PWC), land vegetation condition index (VCI), and the United Nations Food and Agriculture Organization’s (FAO) fish capture data. The results show that SST and SLP displayed significant decadal variability. The decadal variability of sea surface temperature anomalies (SSTA) associated with sea level pressure anomalies (SLPA) has an influence on the land vegetation moisture condition (VCI). Positive SSTA tends to be associated with negative SLPA, and vice versa, in the corresponding ocean areas and most land areas. Consequently, clearly opposing distributions of SSTA and SLPA are observed in the periods 1982–1988 and 1999–2008. With positive SSTA and negative SLPA, VCI tends to increase in value representing more favourable vegetation conditions. Negative SSTA and positive SLPA is generally unfavourable for global vegetation development. The decadal variability of SSTA is closely related to the number of fish species (NFS) doing better or worse based on normalized fish landing data. However, the fishery responses show different yet consistent trends in the three ocean basins. When SSTA is negative, it appears more beneficial for the number of fish species with improved landings in the Atlantic Ocean. However, positive SSTA leads to more fish species with improved landings in the Indian and Pacific Oceans.     

Depth resolved detection of oceanographic variables in the St. Lawrence estuary using a laser fluorosensor: instrument characteristics and first results†

A 64-channel Laser Environmental Airborne Fluorosensor (LEAF) system for sea surface oil detection was modified for continuous depth-resolved detection of gelbstolT and chlorophyll a spectral fluorescence, and water Raman scattering. The system was successfully tested during shipborne and airborne experiments in the St. Lawrence Estuary. System performance was studied, and the best integration time to optimize the signal to noise ratio (S/N) was determined using dilTerent statistical procedures such as the geophysical and robust statistical methods. In general, raw data indicated S/N of about 150, 10, and 4 for Raman. gelbstolT and chlorophyll a signals, respectively. Results with calibrated LEAF spectra clearly indicated that, with this system, water Raman scattering, geibstolT and chlorophyll a signals could be detected down to a maximum depth of 4·8 m. Spatial distribution of these quantities compared well with simultaneously observed ill Silll structure of oceanographic variables, such as, underwater light attenuation, salinity and chlorophyll a fluorescence. In the perspective of modelling primary production in coastal and estuarine waters in the St. Lawrence system (Case II waters), the utilization of LEAF should provide an adequate representation of the spatial and temporal variation pattern of oceanographic variables, at scales between those from ships and satellites.     

On some aspects of precipitation over the tropical Indian Ocean using satellite data

The annual and inter‐annual variability of precipitation over the tropical Indian Ocean is studied for the period 1979–1997, using satellite data from a variety of sensors. The Climate Prediction Center Merged Analysis Precipitation (CMAP), Microwave Sounding Unit (MSU) estimates of rainfall had better correlation with the island rainfall data than the National Centers for Environmental Prediction/National Center for Atmospheric Research Reanalysis (NRA) estimates. A comparison of the mean annual rainfall by different estimates (CMAP, MSU, NRA and GPCP (Global Precipitation Climatology Programme)) showed significant differences with the CMAP, GPCP and MSU estimates depicting maximum off the Indonesian Islands whilst the NRA exhibited maximum in the southern part of the Bay of Bengal and equatorial Indian Ocean. A study of the inter‐annual variability of the monsoon rainfall using the monthly CMAP data over the tropical Indian Ocean for different study areas, namely, Arabian Sea (AS), Bay of Bengal (BB), south Indian Ocean (SIO) and Indian Ocean (IO) showed significant differences during deficit years (1979, 1982, 1986 and 1987), excess monsoon years (1983 and 1988) and also during El Nino Southern Oscillation (ENSO) years (1982, 1987, 1992 and 1997). An analysis of the rainfall anomalies showed positive and negative anomalies in the north‐eastern Bay of Bengal during the summer season of deficit (1986) and excess (1988) monsoon years respectively, whilst the eastern equatorial Indian Ocean showed large positive and negative rainfall anomalies during the autumn season of El Niño years, 1987 (deficit monsoon) and 1997 (normal monsoon) respectively.