Seasonal and Interannual Changes of Indian Oil Sardine, Sardinella longiceps Landings in the Governorate of Muscat (the Sea of Oman)

SERGEY A. PIONTKOVSKI, HAMED S. AL-OUFI, and SAUD AL-JUFAILI

Introduction Given the high abundance of fi sh fi sh, including the Indian oil sar- in Omani waters and its importance dine, Sardinella longiceps, occur in Oman is one of the most impor- to the livelihood of thousands of peo- large quantities (Haleem et al., 2011). tant countries engaged in fi shing in ple, the fi sheries sector is a signifi - Small pelagic stocks in Oman waters the Middle East. The 3,240 km coast- cant sector in the Omani economy. appear to be sustainably fi shed at the line, with a commercial fi shing area of There is a strong fi shing tradition in moment, and research surveys suggest 350,000 km2, has rich fi shing grounds, Oman, and a large number of small that there is room for some further ex- the potential of which has yet to be villages scattered along the coast, pansion in the Arabian Sea. These spe- fully evaluated. A 200 nmi exclusive from which, in 2011, around 40,161 cies form an important component of economic zone extending seaward fi shermen were directly employed in the marine food web because they from the baseline from which the the fi sheries sector operating 18,731 comprise the bulk of the forage for territorial waters are determined, has fi shing boats of which 96% were fi - large fi sh and other predators (Al-Bar- been declared. Omani fi sheries may berglass, 8–10 m in length (Fishery wani et al., 1989). They also contrib- be divided into two broad categories, Statistics Book, 2011). ute signifi cantly to the Omani marine traditional (artisanal) and commer- Small pelagic fi sh are an important fi shery: in 2011 about 27,931 metric cial, with the former representing the component of the artisanal fi shery in tons (t) of sardines were landed along cornerstone of the national industry Oman (Al-Barwani et al., 1989). In the coast of Oman valued at 5.8 mil- and accounting for 96% of landings some regions, coastal pelagic fi sh con- lion RO (~15 million USD). (Fishery Statistics Book, 2011). tribute directly to human consumption Sardines are exploited primarily of fresh, dried, canned, smoked, or fro- with beach seines, cast nets, and gill zen fi sh, hence a large proportion of nets. Small-scale sardine purse sein- protein and nutrition needed for poor ing is already conducted in some lo- Sergey A. Piontkovski (corresponding author, [email protected]) and Saud Al-Jufaili communities, as well as providing in- cations (in particular in south and are with the Department of Marine Science and come for fi shermen. Purse seining is north Al-Batinah regions) along the Fisheries, Sultan Qaboos University, P.O. Box 34, Al-Khod 123, Sultanate of Oman. Hamed a major fi shing method for these spe- Sea of Oman as a result of a recent S. Al-Oufi is with the Ministry of Agriculture cies, including large-scale fi shing for modifi cation by adding rings and a and Fisheries Wealth, P.O. Box 427, Muscat 100, conversion of fi sh into fi sh meal and drawstring to gillnets, thereby form- Sultanate of Oman. oil (particularly by Peru and Chile). ing encircling gill nets. Because they doi: dx.doi.org/10.7755/MFR.76.3.3 In Omani waters, small pelagic require little investment in manpower and equipment, and are effi cient in catching pelagic fi sh, encircling gill- nets are widely used by traditional ABSTRACT—Monthly data on Mus- variability in sardine landings might be fi shermen. Due to sardine schooling cat’s landings of the Indian oil sardine, approximated by the seasonal variations behavior (Misund et al., 2003), beach Sardinella longiceps, along with 23 envi- of the zonal component of wind speed and seines, purse seine, and encircling ronmental parameters ( sea surface tem- chlorophyll-a concentration in the coastal perature, temperature of the mixed layer, and open-sea regions. In terms of interan- gillnets are an effi cient and effective wind speed, kinetic energy of mesoscale nual changes, sardine landings exhibited a method of harvesting. Schooling can eddies, concentration of nitrates, dissolved declining trend from 2001 to 2011 (the time also be facilitated through fi sh ag- oxygen, chlorophyll-a, abundance of phyto- covered by the most complete data set). gregating devices (FAD’s) emitting plankton, zooplankton, and several others) Rising sea temperature, thermal stratifi ca- were analyzed for the period 1994–2011. tion of the water column, and the trophic strong light at night (Fréon and Dago- Seasonal changes were associated with pressure imposed on sardine populations rn, 2000). the time of the winter (Northeast) mon- by large pelagic predators (talang Fishmeal and oil are essential in- soon, with maximal landings in Febru- queenfi sh, Scomberoides commersonnia- gredients in feeds used by the aqua- ary. The multiple regression analysis of nus; kingfi sh, Scomberomorus commerson; culture industry, which grew at an the statistically signifi cant variables se- longtail tuna, Thunnus tonggol; and some lected through the Principal Component others) might be the factors mediating this average rate of 8–20% per year. If Analysis has implied that 51% of seasonal trend. this trend continues, the demand and

50 Marine Fisheries Review prices for fishmeal are likely to in- region (Haleem et al., 2011). On the Table 1.—Common and scientifi c names of the large crease in the future, noting that this other hand, sardines are exposed to pelagic species. will also fl uctuate at decadal scales the press of diverse predators repre- Common names Species because of the fl uctuations in supply, sented by large pelagic species, such Sawtooth barracuda Sphyraena putnamiae Blacktail barracuda Sphyraena qenie especially from the major producers. as the longtail tuna, Thunnus tong- Pickhandle barracuda Sphyraena jello In Oman, due to the decline in the gol; sawtooth barracuda, Sphyraena Sharpfi n barracuda Sphyraena acutipinnis Yellowtail barracuda Sphyraena fl avicauda catch of large pelagics, human con- putnamae; kingfi sh, Scomberomorus Striped bonito Sarda orientalis Sailfi sh Istiophorus platypterus sumption of sardines has increased, commerson; queenfi sh, Scomberoides Frigate tuna Auxis thzard thzard resulting in increased market prices. commersonnianus; Indo-Pacifi c sail- Longtail tuna Thunnus tonggol Yellowfi n tuna Thunnus albacares These issues may result in over-ex- fi sh, Istiophorus platypterus; and some Kingfi sh Scomberomorus commerson ploitation of sardines, possibly endan- other species. Using statistical analysis Queenfi sh Scomberoides commersonnianus gering the stocks in the future. of environmental variables, we have In 2011 and 2012, Oman exported attempted to estimate the trends and about 76% of its small pelagic catch. factors driving seasonal and interan- Data on sardine landings were com- The biggest market for all species is nual changes in sardine landings in the plemented by the monthly frequency United Arab Emirates followed by governorate of Muscat. of fi sh kill incidents and algal blooms Saudi Arabia, Qatar, and Thailand. along the Omani coast, available from Materials and Methods Regionally, sardines contribute over the archives of the Marine Science and 50% to total landings reported for Sardine landings are routinely moni- Fisheries Research Center (Muscat, the western and eastern sides of the tored by the Department of Fisheries Oman) and Al-Gheilani et al. (2012). Arabian Sea (Samuel, 1967; Dhulked Statistics based on a sampling sys- Zonal and meridional components et al., 1982; Fishery Statistics Book, tem established by the Oman-Amer- of wind speed were retrieved from 2011). Along the coast of Oman, ap- ican USAID project (Mathews et al., the NCAR/NCEP reanalysis database proximately 80% of sardine landings 2001). Monthly landings data for sar- (Kistler et al., 2001), in which the from the traditional fi shery are con- dines and the other major fi sh popula- daily averages of wind speed at 10 m tributed by the Indian oil sardine (Al- tions in the Sea of Oman are available above sea level were extracted for the Abdessalam, 1995). from the annual reports published by Sea of Oman region with coordinates An attempt to explain interannu- the Ministry of Agriculture and Fish- lat. 22.5–25.0oN; long. 57.5–60.0oE. al variations of sardine landings on eries Wealth (Fishery Statistics Book, The meridional component has a posi- the basis of fl uctuating environmen- 2009). Common names of the large tive value when the wind is blowing tal characteristics has a 100 year his- pelagic species represented in these re- from south to north. A south wind tory (Hornell, 1910; Longhurst and ports correspond to the names given in has a positive meridional constituent, Wooster, 1990; Kawasaki, 1991; Al- Table 1. while a north wind has a negative one. Jufaili, 2007; George et al., 2012). To assess the relationship between In case of zonal component, values The approach stems from the ecology sardine landings and environmental are positive when the wind is blowing of sardines exhibiting rapid turnover variables, a set of monthly time series from west to east. Thus, a west wind rates of populations (with a doubling was assembled. These time series were has a positive zonal constituent while time of about 15 months), as well as selected to characterize coastal and an east wind has a negative one. their vulnerability to changes of ther- open sea waters. The coastal time se- For the assembly of historical data mal and productive properties of ries were represented by stations BK on temperature in the Sea of Oman, habitats. and F sampled by the Sultan Qaboos we compiled the database on 7,000 In following Longhurst and Wooster University team of researchers over vertical profi les resulting from on- (1990) and George et al. (2012), we the past 8 years (Al-Azri et al., 2009). board casts carried out from 1950 to hypothesized that in the case of the These stations (BK, lat. 23.51oN, long. 2009. These data, were selected to es- Omani traditional fi shery, sardine 58.72oE, and F, lat. 23.67oN, long. timate monthly averaged temperatures landings are representative of stock 58.5oE) are located in the Muscat in the upper 20 m layer for the open abundance. The assumption might be region, 25 km apart along the coast, sea region in the central part of the adequate if the number of fi shing boats with an average depth of 10 m and Sea of Oman (lat. 23–26oN, long. 57– in the region does not increase. In our 20 m, respectively. Monthly data on 60oE). All of the other remotely sensed analysis (carried out in very general the concentration of nitrates, dissolved variables were extracted for the same terms), we also hypothesized that on oxygen, chlorophyll-a, abundance of region. one hand the phytoplankton biomass phytoplankton (total diatoms and di- Satellite derived (4 km and 9 km (assessed by chlorophyll-a concentra- nofl agellates), and zooplankton (cope- spatial resolution SeaWIFS and MO- tion) will characterize the availabil- pods) obtained at these stations were DIS Aqua) monthly Level-3 prod- ity of food for sardines, known to be used for the Principal Component ucts for sea surface temperature and the phytoplankton consumers in the Analysis (Jolliffe, 2002). chlorophyll-a concentration are avail-

76(3) 51 Figure 1.—The map of Oman coast with governorates (left) and averaged seasonal cycles of sardine landings over these governor- ates (2000–11) (right).

able from the National Aeronautics (lat. 18–25ºN, long. 58–62ºE), for the series. Even the time range from 1994 and Space Administration (NASA) period 1997–2009. The methodology to 2000 had data missing, so the con- Ocean Color Group (http://oceancol- of calculations of eddy kinetic energy tinuous seasonal pattern was analyzed or.gsfc.nasa.gov). Monthly time se- per unit mass was that used by Sharma predominantly for the time range from ries of chlorophyll-a and sea surface et al. (1999). 2000 to 2011 (Fig. 2, 3). temperature were acquired for the pe- In the seasonal cycle of landings, Results riod 1997–2011 using the GES-DISC variations were most pronounced in Interactive Online Visualization and Among the governorates of sardine 2000–02, when seasonal gradient be- Analysis Infrastructure software as fi shery in the Sea of Oman, Mus- tween winter and summer values part of the NASA’s Goddard Earth cat and Al-Batinah lead the land- reached one order of magnitude (Fig. Sciences Data and Information Ser- ings. Unfortunately, no plankton and 2). From 2002 to 2009, the seasonal vices Center. physical-chemical sampling has been gradient declined, as did the annual Since mesoscale eddies markedly carried out in Al-Batinah so we were landings. affect monthly variations in chloro- focused on the data available for Mus- To evaluate the components com- phyll-a sea surface temperature, and cat. Also, this is the region with pro- prising the variance of landings, a the other characteristics of the upper nounced seasonal fl uctuations (Fig. spectral analysis was employed (Fig. layer (Piontkovski et al., 2012b), their 1). The maximum matches the time of 4). This statistical procedure enables energetic characteristics were calcu- winter monsoon lasting from Decem- one to break down the total time series lated. The sea surface height anoma- ber to March. In all the other Sea of variance into characteristic compo- lies (required to estimate the kinetic Oman regions maximal landings coin- nents with their own periods. energy of eddies) were produced from cide with the winter monsoon as well. In transforming data for the analy- TOPEX/Poseidon, Jason-1 and Ja- The seasonality of landings with dom- sis, the Tukey smooth window with son-2 altimeter data and acquired from inant winter maximum is the pattern 0.36, 0.24, 0.45, 0.24, and 0.04 Ham- the Archiving, Validation and Inter- pronounced throughout the years. The ming weights was used. One could no- pretation of Satellite Oceanographic reports of the Ministry of Agriculture tice three major peaks contributing to data center website (http://www.aviso. and Fisheries Wealth enabled us to an- the total spectral density of the sardine oceanobs.com). The assessment of the alyze this pattern retrieved from regu- landing variance. The fi rst peak could kinetic energy of eddies was based lar (monthly) data from 2000 to 2011. not be treated as statistically reliable on the altimeter-derived sea surface Earlier data are fragmented and did because the period corresponding to heights for the western Arabian Sea not allow us to construct monthly time this peak is close to the duration of the

52 Marine Fisheries Review Figure 2.—Seasonal cycles of sardine landings in the gov- Figure 3.—Long-term monthly fl uctuations of sardine ernorate of Muscat. landings in the governorate of Muscat.

Figure 4. —The power spectrum of sardine landings in the Figure 5.—Monthly time series of chlorophyll-a con- governorate Muscat. centration (from MODIS-Aqua) and sardine landings in the governorate of Muscat. time series (15 year). Among the other feed on phytoplankton (Haleem et al., In comparing monthly time series of peaks, the annual periodicity seems to 2011). With this regard we used the sardine landings and remotely sensed be the major contributor, followed by chlorophyll-a concentration as the in- chlorophyll-a, we retrieved data from the semi-annual variations. As it was dicator of phytoplankton biomass. the SeaWIFS and MODIS-Aqua ar- shown in Fig. 1, the seasonal changes Our previous studies implied that re- chives. The time series for the later are mostly associated with the time of motely sensed chlorophyll-a (from the one is exemplifi ed in Fig. 5. winter monsoon during which sardine SeaWIFS scanner) is statistically cor- One could notice a certain agree- landings (in February) markedly ex- related with the chlorophyll-a from ment in monthly fl uctuations of chlo- ceeded all the others. the coastal time series of the Mus- rophyll and sardine landings which In Omani waters, sardines mostly cat region (Piontkovski et al., 2011). might be represented in a quantitative

76(3) 53 Figure 6.—Scatterplots of sardine landings vs. chlorophyll-o a concentrationo estimated for two chlorophyll datasets (a: SeaWIFS and b: MODIS scanners) for the Sea of Oman (lat. 23-25 N; long. 58-60 E). form as indicated in the scatter plots Monthly time series of chlorophyll- for sardines (Fig. 8). Obviously, all the of landings versus chlorophyll con- a and sardine landings both exhibited relationships evaluated are negative; centration (Fig. 6). The scatter plots a tendency to decline over years, from they all imply the tendency of sardine characterize the agreement in a very 2002 to 2011 (Fig. 3, 5). In fact, the landings to decrease when landings of general form, averaged over vari- tendency might be traced back to 2001 sardine predators are going up. ous scales of parameter variations. (Fig. 3). As for the earlier data, they To estimate a potential role of the The breakdown of covariance of two are fragmented which does not allow other environmental factors contrib- variables over these scales might be them to be the part of the time series uting to seasonal variance of sardine assessed through the coherency spec- we analyzed. landings we analyzed 23 environmen- trum (Fig. 7). In a given case, the The other potentially important fac- tal variables (Table 2) featuring the coherency spectrum implies the ma- tor is the trophic pressure imposed on coastal and open-sea regions. The se- jor periods (or frequencies) at which sardines by large pelagic predators. lection of variables, based on previous both parameters exhibit maximal cor- We analyzed annually averaged histor- investigations, reported a marked role relation. Obviously, these periods are ical data on landings of large pelagic of these variables in physical-biolog- the annual and semiannual. species acting as potential predators ical coupling in the Arabian Sea and elsewhere (Madhupratap et al., 1994; Logerwell et al., 2001; Logerwell and Smith, 2001; Gaol et al., 2004; Hal- eem et al., 2011; George et al., 2012). The variables were log-transformed and subjected to the Principal Com- ponent Analysis (PCA), which is the data compression procedure enabled to reduce the number of variables (used to describe the variability of data) to a few Principal Components (Factors) refl ecting the compression result (Jol- liffe, 2002). Since seasonal periodicity is one of the most pronounced signals in all the above variables, data for the PCA were arranged in the form of a seasonal cy- cle (from January to December), aver- Figure 7.—The coherency spectrum: sardine landings vs. aged from 2004 to 2008. This was the chlorophyll-a concentration (from SeaWIFS). time range with minimal number of

54 Marine Fisheries Review missed data for all the variables listed. the Muscat region. It should be noted, gradually mediated by summer mon- For the assessment of factor loadings, however, that seasonal patterns of sar- soon) and the reaction of sardine the Varimax normalized matrix was dine landings are distinctly different shoals to the increased abundance of applied to the logarithmically trans- along the Omani coast (Fig. 1), and the diatoms and zooplankton biomass formed variables listed in Table 2. We same goes to the seasonal patterns of was reported for the Kerala coast of constrained the PCA analysis by the chlorophyll-a. This means that a series India. With the monsoon and coastal extraction of three Factors (compo- of regional models on sardine landings upwelling both progressing north- nents), which explained 76% of the to- should be worked out in the future. ward, the sardine shoals followed this tal variance in the system of selected trend to stay in the zones with maxi- Discussion variables. In Table 3, variables with mal food for larvae (Madhupratap et statistically signifi cant loadings are In the Sea of Oman, the sardine is al., 1994). given in bold. one of the major fi shery components Being fi lter feeders, sardines are Apparently, all three Factors are dif- which comprised about 84% of the supposed to be tightly related to spa- ferent by the variables driving the fac- total landings of small pelagic fi sh tial-temporal changes of phytoplank- tor load. For instance, Factor 1 is a averaged for 2001–11 (Fishery Statis- ton concentration. The gut content complex of interacting atmospheric tics Book, 2011). We used historical analysis of the sardines caught along and hydrophysical variables; Factor 2 data to describe seasonal and interan- the coast of the Sea of Oman (in the is mainly associated with the interplay nual trends in landings. Apparently, Muscat and Sohar regions) has indi- between three variables signifi cantly in the case of sardine landings, both cated that phytoplankton comprise contributing to this Factor, which is patterns might be interpreted in terms about 62–68% of their diet (Haleem et the zonal component of wind speed, of environmental factors rather than al., 2011). It is not surprising that sea- concentration of chlorophyll-a in the the pressure on fi sh stocks. Seasonal sonal fl uctuations of sardine landings open and coastal waters, and sardine fl uctuations of landings are correlated were statistically associated with the landings. Eventually, Factor 3 is load- with the forcing imposed by monsoon- phytoplankton-related parameter used ed primarily by seasonal variation of al winds on the upper mixed layer of in our study (chlorophyll-a concentra- the concentration of diatoms in coastal the sea (in our case, the zonal compo- tion). In the eastern Arabian Sea, in waters (the Muscat region). nent of the wind speed). The winter coastal waters of India, the remotely A group of variables contributing (northeast) monsoon is the major force sensed chlorophyll-a could explain to Factor 1 has explained 52% of the mediating the seasonality of photosyn- up to 39% of interannual variations total variance. Factor 2 contributed an thetically active radiation, wind speed, in sardine landings (George et al., additional 15%, which came to 67% sea surface temperature, nutrients, and 2012). of the total variance explained. Fac- chlorophyll-a concentration in the re- In some way, seasonal fl uctuations tor 3 has added 9% which raised the gion (Piontkovski et al., 2011). All of of sardine landings and their regional level of explained variance up to 76% these parameters exhibit pronounced differences might be infl uenced by me- (Table 4). Overall, the three principal seasonal patterns in the Sea of Oman soscale eddies propagating through the components (three Factors) employed, and a certain group of environmental coastal waters. In the California Cur- have explained the major part of sea- parameters correlate well with sardine rent, the elevated densities of sardine sonal variation within the system of landings. larvae were associated with cyclonic proposed variables (Table 2). A sub- Similar results were reported for eddies and the peripheries of anticy- sequent stepwise multiple regression the other regions. In the southeastern clonic eddies which are believed to be analysis of the statistically signifi cant part of the Indian Ocean, fl uctuations favorable areas for sardine recruitment variables denoted by the PCA has im- of sardine catches were positively cor- (Logerwell and Smith, 2001; Loger- plied that 51% of seasonal variation related to the remotely sensed chlo- well et al., 2001). Warm eddies could in sardine landings might be approxi- rophyll-a (Gaol et al., 2004). In the often branch out warm streamers vis- mated by the seasonal variations of Mediterranean Sea, a set of remote- ible in satellite images of sea surface the zonal component (east-west) of ly sensed parameters (chlorophyll-a, temperature with streamer structures wind speed and chlorophyll-a con- photosynthetically active radiation, sea of about 10 km in width. Simultane- centration in the coastal and open sea surface temperature, sea level anom- ous satellite measurements and air- regions. aly, and some others) was suffi cient craft observations on sardine schools Predictive properties of the above to explain the location of the juvenile have shown that sardines tend to use multiple regression might be exem- grounds of European sardines. These warm streamers for migration towards plifi ed by the plot of predicted values areas were mostly located inshore and the coast, with the formation of dense of sardine landings as the function of often in proximity to river mouths schools at the head of warm stream- actual reported landings (Fig. 9). The (Schismenou et al., 2008). ers, in which fi shing grounds are of- model is aimed at forecasting sea- Nutrient enrichment of coastal wa- ten formed (Sugimoto and Tameishi, sonal changes of sardine landings in ters through river runoff (which is 1992; Tameishi et al., 1994).

76(3) 55 Figure 8.—The relationship between landings of sardines (t) vs. landings of the sardine predators (1994–2011).

56 Marine Fisheries Review Table 2.—Characteristics of the variables used for the Principal Component Analysis. Table 3.—Results of the Principal Component Anal- ysis: factor loadings (Varimax normalized). Marked Variable loadings are > 0.70. Variables with statistically signifi - No. Variable name abbreviation Data source cant loadings are given in bold.

1 Atmospheric pressure Atm Press NCAR/NCEP database Variable Factor 1 Factor 2 Factor 3 2 Atmospheric temperature Atm Temp NCAR/NCEP database 3 Outgoing low wave radiation OLR NCAR/NCEP database Atm Press 0.60 0.53 -0.50 4 Atmospheric precipitation Precipitat NCAR/NCEP database Atm Temp -0.77 -0.55 0.26 5 Optical thickness of aerosol Aerosol MODIS database OLR -0.82 -0.47 0.20 6 Zonal component of wind speed Z-wind NCAR/NCEP database Precipitat 0.85 0.34 -0.03 7 Meridional component of wind speed M-wind NCAR/NCEP database Aerosol -0.88 0.12 0.24 8 Concentration of dissolved oxygen Oxy-20 SQU Historical database Z-wind 0.09 -0.73 0.27 9 Sea surface temperature SST-MODIS MODIS database M-wind -0.67 -0.35 0.57 10 Temperature (20m deep; averaged SST-MODIS -0.87 -0.43 0.15 for the central part of the Sea of Oman) Temp-20 SQU database Temp 20 -0.82 -0.36 0.15 11 Kinetic energy of mesoscale eddies EKE eddy SQU database EKE eddy 0.92 0.05 -0.01 12 Concentrations of nitrates (coastal NO3-BK 0.81 0.19 -0.03 time series; station BK) NO3-BK SQU database Dino-BK -0.55 -0.40 0.40 13 Concentration of dissolved Diatom-BK -0.27 0.20 0.82 oxygen (20m deep; averaged for the Chl-SeaWIFS 0.40 0.72 -0.24 central part of the Sea of Oman) Oxy-20 SQU database Chl -F+BK 0.09 0.85 0.27 14 Concentration of dissolved Oxy-20 0.91 -0.12 0.00 oxygen (coastal time series, station BK) Oxy-BK SQU database Oxy-BK -0.30 -0.45 0.68 15 Concentration of chlorophyll-a Sardine Land 0.14 0.75 -0.13 (coastal time series; stations F+BK) Chl-F+BK SQU database HABs 0.42 0.52 0.22 16 Concentration of dinofl agellates Fish kills -0.56 0.54 -0.20 (coastal time series; station BK) Dino-BK SQU database Chl-MODIS 0.45 0.68 -0.07 17 Concentration of diatoms (coastal Copepod-BK 0.13 -0.06 0.54 time series; station BK) Diatom-BK SQU database 18 Sardine landings Sardine land Ministry of Fisheries Reports 19 Concentration of chlorophyll-a (from SeaWIFS) Chl-SeaWIFS SeaWIFS database 20 Concentration of chlorophyll-a (from MODIS) Chl-MODIS MODIS database 21 Total copepod abundance (coastal time series; station BK) Copepods-BK SQU database for all the major fi sheries regions along 22 Frequency of algal blooms Algal blooms All Gheiani et al., 2012 the coast, from Musandam, through 23 Frequency of fi sh kill incidents Fish kills All Gheiani et al., 2012 Al-Batinah, to the Muscat (Haleem et al., 2011), which is in fact the Sea of Oman scale. On this scale, rising sea The western Arabian Sea is known encircling gear with rings was enforced temperature and thermal stratifi cation for its vigorous fi eld of mesoscale for many years in the sardine fi shery. of the water column were reported (Pi- eddies that markedly affect phyto- Among the other trends reported, ontkovski et al., 2012a) and might be plankton and zooplankton distribution which might be ecologically important the components partially explaining (Piontkovski and Banse, 2006). As in the analysis of a declining trend in this trend. far as the Sea of Oman is concerned, sardine landings, the frequency of fi sh The declining trend in sardine land- recent studies have shown an interan- kill and algal bloom incidents might ings (from 2001 to 2011) could be a nual increase in seasonal fl uctuations be taken into account. Apparently, fragment of a more prolonged period of the kinetic energy of eddies as well the annual frequency of fi sh kills and of fl uctuations. In the other geographi- as the increase in seasonal fl uctua- coastal algal blooms in Omani coastal cal regions, the 30- to 60-year time se- tions of chlorophyll-a in 2000–08 (Pi- waters had gradually increased in the ries of sardine landings enabled a tight ontkovski et al., 2012b). Perhaps both past 30 years (Al-Azri et al., 2012; Al- coupling to be elucidated, between events were not favorable for sardines, Gheilani et al., 2012). landings and climatic indices featur- so both may have contributed to a de- In the Sea of Oman, over 70% of ing these regions. For instance in the clining trend in sardine landings. This coastal algal blooms are reportedly northwestern Mediterranean Sea, the is consistent with the ranking of the caused by the dinofl agellate, Noctiluca Western Mediterranean Oscillation kinetic energy of eddies as a power- scintillans, which is a species actively index (WeMOi) is a sensitive indica- ful (high scored) variable in Factor 1 avoided by the Indian oil sardine. High tor of climate variability. It was shown of our Principal Component Analysis concentrations of Noctiluca usually that positive WeMOi values were (Table 2). lead to a scarcity of sardines (Nair, correlated with low sea surface tem- As for fi shing pressure and its role 1958). Furthermore, Noctiluca has perature, high river runoff, and high in the formation of declining trend, the not been reported in their stomachs landings per unit of effort. Conversely, number of fi shing boats in the region (Prasad, 1953; Sekharan, 1966). Being did not show steady gradual increase, a heterotrophic organism feeding on Table 4.—Results of the eigenvalue extraction by the Principal Component Analysis: the total and cumula- with the reported number of 1,858 small phytoplankton, Noctiluca acts as tive variance explained by three Factors (Table 2). boats in 2004 vs. 1,624 boats in 2007, the competitor for food with sardines. Eigen- % total % cumulative and 1,946 boats in 2011 for the Muscat It should be emphasized that the Factor value variance variance governorate, for instance. In terms of long-term declining trend in sardine 1 11.34 51.53 51.53 the other factors affecting landings it landings (from 2001 to 2011) is a ba- 2 3.32 15.10 66.64 should be noted that a ban on the use of sin-scale phenomenon. It was reported 3 1.95 8.85 75.48

76(3) 57 od of about 60 years (Baumgartner et al., 1992). Recent estimates of typical fl uctuations of temperature anomalies and fi sh landings in the world ocean report the period of 60–70 years as the most prominent in the past millennium (Klyashtorin and Lyubushin, 2007; Zhen-Shan and Xian, 2007; Lyubushin and Klyashtorin, 2012). The issue of how well the Sea of Oman fi ts this pic- ture is yet to be investigated. Acknowledgment This paper was supported by the Min- istry of Agriculture and Fisheries Wealth (Oman) and The Research Council grant # ORG/EBR/11/002 (Oman). Figure 9.—A scatterplot of predicted vs. actual landings based on multiple regression model (Table 3). Y = 1.38 Literature Cited + 0.52X; R= 0.72. Dashed lines: 95% confi dent interval. Al-Abdessalaam, T. Z. S. 1995. Marine spe- cies of the Sultanate of Oman. Mar. Sci. Fish. Cent., Minist. Agric. Fish., Sultanate of Oman, Pub.46/95, 412 p. Al-Azri, A. R, S. A. Piontkovski, K. A. Al-Hash- negative WeMOi values were associat- along the coast could provide useful mi, J. I. Goes, and H. R. Gomes. 2009. Chlo- ed with high sea surface temperature, information as well. For instance, the rophyll-a as a measure of seasonal coupling low river runoff, and low landings per declining trend in the Sea of Oman between phytoplankton and the monsoon periods in the Gulf of Oman. Aquat. Ecol. unit of effort. This means that the neg- tends to switch to the “no trend situ- 44:449–461. ative WeMOi was the phase with unfa- ation” in the neighboring Al-Sharqiya ______, S. A. Piontkovski, K. A. Al-Hash- mi, H. Al-Gheilani, H. Al-Habsi , S. Al-Khu- vorable environmental conditions and region (which is about 400 km to the saibi, and N. Al-Azri. 2012. The occurrence would decrease the survival, growth, south down the coast). Along the Indi- of harmful algal blooms (HABs) in Omani and reproduction of sardines during an coast stretched for 7,500 km, inter- coastal waters. Aquat. Ecosystem Health Manage. Soc. 15(S1):56–63. their life cycle (Martin et al., 2012). annual trends of sardine landings vary Al-Barwani, M. A. Prabhakar, J. Dorr, and M. The Omani time series of sardine greatly as well, from positive to fl uc- Al-Manthery. 1989. Studies on the biology landings is yet to be suffi cient for tuations with no apparent interannual of Sardinella longiceps (Valenciennes) in the Sultanate of Oman, 1985–1986. Kuwait Bull. that kind of analysis. Nonetheless, a trends (Vivekanandan et al., 2008). Mar. Sci. 10:201–209. simple comparison of our data to the Migrations of sardines along the coast Al-Gheilani, H. M., K. Matsuoka, A. Y. Al-Kin- di, S. Amer, and C. P. Waring. 2012. Fish data characterizing the opposite site are believed to be caused by changes kill incidents and harmful algal blooms in of the Arabian Sea implied that cur- of the preferential temperature zones Omani waters. J. Agric. Mar. Sci. 16:33–46. rent interannual trends of sardine land- and predator press (Armstrong et al., Al-Jufaili, S. 2007. Conceptual model for sar- dine and anchovy inverse cyclic behav- ings might be opposite, although both 1991; O’Donoghue et al., 2010). Both ior in abundance. J. Food Agric. Environ. of them developed on the background a temperature increase in the upper 5:317–327. of increasing temperature in the upper mixed layer over the past 50 years and Armstrong, M. J., P. Chapman, S. F. J. Dudley, I. Hampton, and P. E. Malan. 1991. Occur- mixed layers. For instance, the declin- a pronounced increase in predators af- rence and population structure of pilchard ing trend in sardine landings reported fecting sardine have been reported for Sardinops ocellatus, round herring Etrumeus whiteheadi and anchovy Engraulis capensis for the Sea of Oman might be super- the Sea of Oman (Piontkovski et al., off the east coast of southern Africa. S. Afr. imposed on a rising trend of landings 2012a; Fig. 8). J. Mar. Sci. 11:227–249. along the southwestern coast of India As we noticed, the declining trend Baumgartner, T. R., A. Soutar, and V. Ferrira- Bartina. 1992. Reconstruction of the history (Vivekanandan et al., 2008). Does this might be a fragment of a longer-term of Pacifi c sardine and northern anchovy pop- mean that the Omani sardines migrat- fl uctuation. What are the typical peri- ulations over the past two millennia from sed- ed eastward, to the Indian coasts, or ods of these fl uctuations? Data from iments of the Santa Barbara basin, California. CalCOFI Rep. 33:24–40. have they migrated southward, along the Pacifi c Ocean imply naturally oc- Chavez, F. P., J. Ryan, S. E. Lluch-Cota, and M. the Omani coast? The distances of sar- curring periods in sardine landings of Ñiquen. 2003. From anchovies to sardines and back: multidecadal change in the Pacifi c dine feeding migrations could be thou- about 50 years (Chavez et al., 2003). Ocean. Science 299:217–221. sands of kilometers (Schwartzlose et Reconstruction of the history of Pa- Dhulked, M. H., C. Muthiah, S. Rao, N. S. Ad- al., 1999; Wada and Kashiwai, 1991). cifi c sardine biomass over the past hakrishnan. 1982. The purse seine fi shery of Mangalore (Kamataka). Mar. Fish. Info. Serv. Understanding regional trends of in- two millennia from sediments showed T&E Ser. Cent. Mar. Fish. Res. Inst., Cochin terannual changes in sardine landings that sardines tend to vary over a peri- 37:13–15.

58 Marine Fisheries Review Fishery Statistics Book. 2009. Sultanate of Longhurst, A. R., and W.S. Wooster. 1990. Prasad, R. P. 1953. Swarming of Noctiluca in Oman, Muscat, Minist. Fish. Wealth, 147 p. Abundance of oil Sardine (Sardinella lon- the Palk Bay and its effect on the “Choodai” ______. 2011 Sultanate of Oman, Muscat, giceps) and upwelling on the southwest fi shery with a note on the possible use of Minist. Fish. Wealth, 156 p. coast of India. Can. J. Fish. Aquat. Sci., Noctiluca as an indicator species. Proc. Ind. Fréon, P., and L. Dagorn. 2000. Review of fi sh 47:2407–2419. Acad. Sci. 38:40–47. associative behavior: Toward a generalization Lyubushin, A. A., and L. B. Klyashtorin. 2012. Samuel, C. T. 1967. An analysis of the ma- of the meeting point hypothesis. Rev. Fish. Short term global DT prediction using (60– rine fi sh catch in Kerala from 1957–1958 Biol. Fish. 10:183–207. 70)- years periodicity. Repr. from Energy to 1965–1966. Bull. Dep. Mar. Biol. Ocean- Gaol, J. L., B. P. Pasaribu, D. Manurung, and and Environ., 23(1):75–85. Multi-science ogr. 3:61–67. R. Endriani. 2004. The fl uctuation of chloro- Pub. Co. Ltd. Essex. Schismenou, E., M. Giannoulaki, V. D. Vala- phyll-a concentration derived from satellite Madhupratap, M., S. R. Sheyte, K. N. V. vanis, and S. Somarakis. 2008. Modeling imagery and catch of oily sardine (Sardinella Nair, and S. R. S. Nair. 1994. Oil sardine and predicting potential spawning habitat of emuru) in Bali Strait. Int. J. Remote Sensing and Indian mackerel: their fi shery prob- anchovy (Engraulis encrasicolus) and round Earth Sci. 1:22–34. lems and coastal oceanography. Curr. Sci. sardinella (Sardinella aurita) based on satel- George, G., B. Meenakumari, M. Raman, S. 66:340–348. lite environmental information. Hydrobiolo- Kumar, P. Vethamony, M. Babu, and X. Ver- Martín, P., A. Sabatés, J. Lloret, and J. Martin- gia 203:215–223. lecar. 2012. Remotely sensed chlorophyll: a Vide. 2012. Climate modulation of fi sh popu- Schwartzlose, R. A., J. Alheit, A. Bakun, T. R. putative trophic link for explaining variabil- lations: the role of the Western Mediterranean Baumgartner, R. Cloete, R. J. M., Crawford, ity in Indian oil sardine stocks. J. Coast. Res. Oscillation (WeMO) in sardine (Sardina W. J. Fletcher, Y. Green-Ruiz, E. Hagen, T. 28:105–113. pilchardus) and anchovy (Engraulis encra- Kawasaki, D. Lluch-Belda, S. E. Lluch-Cota, Haleem, S. Z. A., N. Jayabalan, F. Al-Kiyumi, sicolus) production in the north-western Med- A. D. MacCall, Y. Matsuura, M. O. Nevárez- L. Al-Kharusi, S. Al-Habsi, and A. Al-Mar- iterranean. Climatic Change 110:4925–4939. Martínez, R. H. Parrish, C. Roy, R. Serra, K. zouqi. 2011. Fishery, biology and population Mathews, C.P., J. Al-Mamry, and S. Al-Habsy. V. Shust, M. N. Ward, and J. Z. Zuzunaga. dynamics of three small pelagic fi sh species 2001. Precautionary management of Oman’s 1999. Worldwide large-scale fl uctuations of (Indian oil sardine Sardinella longiceps, In- demersal fi shery. In First International Con- sardine and anchovy populations. S. Afr. J. dian mackerel Rastrelliger kanagurta and ference on Fisheries, Aquaculture and Envi- Mar. Sci. 21:289–347. Indian scad Decapterus russeli) from the Sul- ronment in the NW Indian Ocean, p. 41–49. Sekharan, K. V. 1966. On the food of the sar- tanate of Oman. Mar. Sci. Fish. Cent., Minist. Sultan Qaboos University, Sultanate of dines, Sardinella albella (VAL.) and S. gib- Agric. Fish. Wealth, Muscat, Proj. Final Rep., Oman, Muscat. bosa (BLEEK.) of the Mandapam area. Ind. Part II, p. 126–167. Misund O. A., J. C. Coetzee , P. Fréon , M. Gar- J. Fish. 13:96–141. Hornell, J. 1910. Report on the feasibility of op- dener , K. Olsen, I. Svellingen, and I. Hamp- Sharma, R., A. K. S. Gopalan, and M. M. Ali. erating deep-sea fi shing boats on the coasts of the Madras Presidency with special reference ton. 2003. Schooling behaviour of sardine 1999. Interannual variation of eddy kinetic to the selection of Fishing Centres and Har- Sardinops sagax in False Bay, South Africa. energy from TOPEX altimeter observations. bours of Refuge. Madras Fish. Bull. 4:33–70. Afr. J. Mar. Sci., 25:185–193. Mar. Geodesy 22:239–248. Jolliffe, I. T. 2002. Principal component analy- Nair, R. V. 1958. The sardines. In S. Jones Sugimoto, T., and H. Tameishi. 1992. Warm- sis. Springer Ser. Stat., Springer-Verlag, N.Y., (Editor), Fisheries of the west coast of In- core rings, streamers and their role on the p. 1–487. dia, p. 31–37. Cent. Mar. Fish. Res. Inst. fi shing ground formation around Japan. Kawasaki, T. 1991. Long-term variability in the Mandapam Camp. Deep-Sea Res. Pt. A 39:S183–S201. pelagic fi sh population. In T. Kawasaki, S. O’Donoghue, S. H., L. Drapeau, S. F. J. Dudley, Tameishi, H., Y. Naramura, and H. Shi- Tanaka, Y. Tola, and A. Taniquchi (Editors), and V. M. Peddemors. 2010. The KwaZulu- nomiya. 1994. Role of warm streamers in The long-term variability of pelagic fi sh pop- Natal sardine run: shoal distribution in rela- the northward migration of Japanese sar- ulations and their environment, p. 47–60. Per- tion to nearshore environmental conditions, dine (Sardinops melanostictus) off Sanriku gamon Press, N.Y. 1997–2007. Afr. J. Mar. Sci. 32:293–307. (Japan). Bull. Japan. Soc. Sci. Fish. 60:45–50. Kistler, R., E. Kalnay, W. Collins, S. Saha, Piontkovski, S. A., and K. Banse. 2006. Over- Vivekanandan, E., N. G .K. Pillai, and M. Raja- G.White, J. Woollen, M. Chelliah, W. Dool, view of results. In K. Banse and S. A. Pi- gopalan. 2008. Adaptation of the oil sardine R. Jenne, and M. Fiorino. 2001. The NCEP- ontkovski (Editors), The mesoscale structure Sardinella longiceps to seawater warming NCAR 50-Year reanalysis: monthly means of the epipelagic ecosystem of the open along the indian coast. In P. Natarjan, K.V. CD-ROM and documentation. Bull. Am. Me- northern Arabian Sea, p. 1–9. Univ. Press, Jayachandran, and S. Kannaiyan (Editors), teorol. Soc. 82:247–268. Hyderabad. Glimpses of aquatic biodiversity, p.111-119. Klyashtorin, L. B., and A. A. Lyubushin. 2007. ______, A. R. Al-Azri, and K. A. Al- Rajiv Gandhi Chair Spec. Publ. 7, Cochin Cyclic climate change and fi sh productivity. Hashmi. 2011. Seasonal and interannual vari- Univ. Sci. Technol. Kochi. In G.D. Sharp (Editor), p. 1–158.VNIRO, ability of chlorophyll a in the Gulf of Oman Wada, T., and M. Kashiwai. 1991. Changes in Moscow. compared to the open Arabian Sea regions. growth and feeding ground of Japanese sar- Logerwell, E., B. Lavaniegos, and P. Smith. Int. J. Remote Sensing 32:7703–7715. dine with fl uctuation in stock abundance. In 2001. Spatially-explicit bioenergetics of ______, H. Al-Gheilani, B. Jupp, A. R. K. Kawasaki, S. Tanaka, Y. Toba, and A. Tani- Pacifi c sardine in the Southern California Al-Azri, K. A. Al-Hashmi. 2012a. Interan- guchi (Editors), Long-term variability of pe- Bight: are mesoscale eddies areas of excep- nual changes in the Sea of Oman ecosystem. lagic fi sh populations and their environment, tional prerecruit production? Progress Ocean- Open Mar. Biol. J. 6:38–52. p. 181–190. Pergamon Press, N.Y. ogr. 49:391–406. ______, N. Nezlin, A. R. Al-Azri, and Zhen-Shan, L., and S. Xian. 2007. Multi-scale ______and P. Smith. 2001. Mesoscale ed- K. A. Al-Hashmi. 2012b. Mesoscale eddies analysis of global temperature changes and dies and survival of late stage Pacifi c sardine and interannual trends of physical-biological trend of a drop in temperature in the next (Sardinops sagax) larvae. Fish. Oceanogr. coupling in the Sea of Oman. Int. J. Remote 20 years. Meteorol. Atmos. Physics 10:13–25. Sensing 33:5341–5346. 95:115–121.

76(3) 59 60 Marine Fisheries Review