Long-term Changes in Sea Surface Temperature at Selected Locations in the Sea of Oman and the Arabian Sea off Oman
2. NOAA Pacific Marine Environmental Laboratory, 7600 Sandpoint Way NE, Seattle, WA 98115, USA
Author Correspondence author
International Journal of Marine Science, 2013, Vol. 3, No. 18 doi: 10.5376/ijms.2013.03.0018
Received: 08 Apr., 2013 Accepted: 19 Apr., 2013 Published: 22 Apr., 2013
Sarma et al., 2013, Long-term Changes in Sea Surface Temperature at Selected Locations in the Sea of Oman and the Arabian Sea off Oman, International Journal of Marine Science, Vol.3, No.18 145-150 (doi: 10.5376/ijms.2013.03.0018)
Long-term changes in the sea surface temperature (SST) at two locations off Oman were investigated using Hadley Center SST for the period 1961~2009. A mean annual increase in SST by 0.32℃ was noticed in the Sea of Oman while an increase of 0.53℃ was noticed in the western Arabian Sea. The shift in SST is higher off Muscat than off Masirah during the study period. The bi-modal variability of the SST in the study region is successfully simulated by a cyclic model developed utilizing SST data for 1961~2009 period. An increase in summer warming and decrease in winter cooling are evident in the annual SST distribution at both the locations. The decadal variability off Masirah in the western Arabian Sea showed that the standard deviation of SST switched its character post-1990. The SST variance in Sea of Oman showed a decadal-scale change but in western Arabian Sea, it was nearly unchanged until 1990 and rapidly declined post-1990 period. The large shifts in SST apparently caused higher variability in the sea surface height (SSH) anomalies post-1990 period.
Sea surface temperature (SST) is an important oceanographic variable as it provides the basis for sea level and climate change studies. Regime shifts in SST are noticed worldwide and are characterized by an abrupt transition from one quasi-steady climatic state to another. The transition period being much shorter than the lengths of the individual epochs of each climatic state and tend to impact the other dynamics such as biological processes (Yasunaka and Hanawa, 2002; 2005). The seasonal nature of the SST in the Arabian Sea is bi-modal while in the Sea of Oman, it is unclear, but reported to be largely uni-modal (Colborn, 1971). The Indian Ocean experienced anomalous warming in 1976~1977 indicating an abrupt shift in climate (Nitta and Yamada, 1989; Aoki et al., 2003; Terray and Dominiak, 2005). Khan et al (2004, 2008) reported an increasing trend of SST in the Arabian Sea over the period 1985~1998 and argued that the northern coast of Oman in the Sea of Oman and the east coast of Oman in the northwestern Arabian Sea exhibited different time scales of inter-annual and seasonal variability of the SST. Anomalous and persistent increase in SST during early summer by about 2℃ off Masirah Island as well as in the Sea of Oman was noticed before the severe cyclones “Gonu” in 2007, and “Phet” in 2010 made landfall along the east coast of Oman (Sarma, 2011).
Long-term trends in SST variability in the Sea of Oman and the northwestern Arabian Sea along Oman are largely under-investigated till date. In this study, we investigated the SST variability at two locations along Oman, which have different warming and cooling patterns. A cyclic model was developed to simulate the bi-modal nature of the seasonal SST changes and the long-term trend in SST at the two selected locations. Further, the SST anomalies (SSTA) at the study locations are compared with sea surface height anomalies (SSHA) to examine the impact of the SST changes on sea level changes at the selected locations.
Sea surface temperature (SST) over a one degree-square at two locations; Muscat (23~24°N, 59~60°E) in Sea of Oman and Masirah Island (19~20°N, 58~59°E) in the Arabian Sea (Figure 1) were extracted from the Hadley Center's SST data for the period 1961 to 2009. HadISST1 temperatures are reconstructed using a two-stage reduced-space optimal interpolation procedure, followed by superposition of quality-improved gridded observations onto the reconstructions to restore local detail (Rayner et al., 2003). The SST anomalies (SSTA) are calculated as SST minus mean SST of the entire period (1961~2009).
Figure 1 Study region and the areas selected (1×1 degree boxes) for present analysis |
Monthly average SST for the years 1961 to 2009 was modeled using the following equation and was performed separately for each location:
Where T is the predicted monthly average SST, A1, P1 and α1 are the amplitude, period and phase shifts, respectively of the first mode and A2, P2 and α2 are the amplitude, period and phase shift of the second mode. The slope and the y-intercept (C) are the linear parameters that describe the trendline of the whole data set. The latter constants were estimated prior to estimating the modal parameters and all parameters were estimated by minimizing the negative log likelihood assuming that the errors were normally distributed with mean zero and variance σ2. The likelihood profile technique (Lebreton et al., 1992) was used for estimating 95% confidence intervals (CI). The slopes of the linear trendlines were tested to determine if they were significantly different from zero (Zar, 1999).
The sea surface temperature for the period from 1961 to 2010 gave a mean SST that is significantly higher in Muscat than in Masirah (t-stat0.05, DF=1174=7.48; P=7.5×10-14). Showing interannual and longer time fluctuations in SST (Table 1), SST is noticeably higher post-1985 at Masirah.
Table 1 Statistics of annual SST off Muscat and Masirah during the study period |
The SST anomaly (SSTA) from 1961 to 2009 showed a distinct shift in the SST distribution after 1984 (Figure 2). A quantum jump of about 0.5℃ is seen in the SSTA post-1984 at both locations. At both locations, there was a positive increase in SST over the period of the study and for both locations the slopes were significantly different from zero (Masirah: t-stat0.05, DF=586=3.88, P=0.0001; Muscat: t-stat0.05, DF=586=2.50, P=0.01). The rate of increase of temperature over the study period was higher in Muscat than Masirah (larger slope value). Muscat shows two prominent peaks with the most prominent one occurring every 12 months while the smaller mode occurring approximately every 6 months (Table 2). The difference between the amplitudes is 0.32℃ indicating that the two modal peaks are similar in amplitude. The most prominent peak occurs annually during the month of July and the smaller peak during May. In the case for Masirah, though bi-modal in nature it has its prominent peak occurs every 6 months while less prominent peak occurring every 12 months. The amplitude difference between these peaks is 1.48℃ and shows a large difference in amplitude prominence between them. In Masirah, the prominent peak occurs in the month of November while the less prominent one occurs in the February-March period. Confidence intervals for the P1 parameter could not be estimated for the Masirah location as the maximum likelihood surface for this parameter was flat around the global minimum.
Figure 2 Distribution of SSTA for Muscat (red line) and Masirah (blue line) for the years 1961 to 2010 |
Table 2 Model parameter estimates and their 95% confidence for the selected locations, Muscat and Masirah |
Figure 3 gives the variance estimates of SST grouped in 10-year intervals for the complete fitted models. There are two general trends in the variance, firstly, in Muscat the variance for the 10-year groups is fairly constant for the period 1961~2009 while for Masirah there has been a rapid decrease in variance since 1981~1990. Secondly, between the years 1961 to 1990 variance estimated for the Masirah Island location always exceeded that of Muscat but from 1991 the variance of Muscat has exceeded that of Masirah Island and more noticeably so in the 2001 to 2009 period. The intra-annual distribution of SST variance shows that the low or high variance at the study locations occurs nearly simultaneously (Figure 4). The low variance off Masirah occurred during February-March (0.11℃) and of high variance is July (0.88℃). Similarly, lower SST variance (0.11℃) off Muscat occurred in February and October while higher variance (0.24℃) occurred in April and August. The SST variance off Masirah is large during summer (June~August) with a peak variance of 0.9℃ during peak summer period (July) and dramatically reduced to <0.2℃ during winter. The inter-annual standard deviations and variances have dramatically increased from pre-1984 to post-1984 period at both locations. The monthly mean SST showed significant difference in the pattern of variability between the selected locations as indicated by the slope (0.0115℃ off Muscat and 0.0225℃ off Masirah). Table 2 shows the amplitude, period and phase of the SST at the study locations estimated from the model. The amplitude of primary warming is higher off Masirah (2.25℃) compared to Muscat (1.55℃).
Figure 3 Decadal distribution of variance (℃) of SST at the study locations for the period 1961 to 2009 |
Figure 4 Monthly distribution of variance (℃) of SST at the study locations for the period 1961 to 2009 |
The bi-modal model parameter estimates given in Table 2 are utilized to simulate the SST changes at the study locations and the model fits are shown in Figure 5 and Figure 6. The annual and semi-annual changes in the SST were better simulated off Muscat where the standard deviation is relatively smaller. The model simulation shows very small departures from the cooling phase off Masirah in certain years. The impact of long-term warming on sea level was examined by comparing SSTA with SSHA from altimetry for the period 1993 to 2009. The SSHA (Figure 7) exhibited, in general, a continued rising trend between 1993 and 2009 following the similar trend in SSTA. The sea level variability occurred simultaneously in the Sea of Oman and in the western Arabian Sea along the east coast of Oman during the above period. However, the variability is larger in the Sea of Oman compared to the western Arabian Sea. While the sea level was gradually rising, it also exhibited an embedded 6-year cycle of rise and fall in western Arabian Sea off Masirah, but such signal was not clear in Sea of Oman.
Figure 5 Modeled SST (solid line) versus observed SST (dots) off Muscat for the period 1961 to 2009 |
Figure 6 Modeled SST (solid line) versus observed SST (dots) off Masirah for the period 1961 to 2009 |
Figure 7 Annual mean SSH anomalies (cm) from altimetry for Muscat (red line) and for Masirah (blue line) |
4 Discussion
Al-Gheilani H.M., Matuoka K., Al-Kindi A.Y., Amer S., and Waring C., 2011, Fish kill incidents and harmful algal blooms in Omani waters, Journal of Agriculture and Marine Science, 16: 23-33
http://dx.doi.org/10.5376/ijms.2012.02.0005
http://dx.doi.org/10.5376/ijms.2013.03.0008
Zar J.H., 1999, Biostatistical Analysis., Prentice Hall, New Jersey, pp.663
. PDF(277KB)
. FPDF(win)
. HTML
. Online fPDF
Associated material
. Readers' comments
Other articles by authors
. Y.V.B. Sarma
. Anesh Govender
. Ebenezer S. Nyadjro
. Sergey Piontkovski
Related articles
. Sea of Oman
. Western Arabian Sea
. Sea surface temperature
. Sea level changes
. Long-term SST change
Tools
. Email to a friend
. Post a comment