Atmospheric forcing over Holocene productivity in the Alboran Sea (W Mediterranean) from coccolithophores and Mg/Ca-estimated SST records (стр. 3 )

5.2.  Primary productivity and SST general trends

The general trend shown by δ18O and SST (Mg/Ca and alkenone-estimated) is in good accord with the insolation curve (July, 65°N) (Figure 4) (Laskar, 1990). This fact suggests that insolation could have played an important role as forcing long-term mechanism over SST during the Holocene, which in turn had an important effect on δ18O general trend. SST (Mg/Ca and Uk’37-estimated) show few similarities with the NAR variability of G. muellerae and WWG, usually used as SST paleoindicators (cold and warm respectively) (McIntyre and Bй, 1967; Okada and Honjo, 1973; Weaver and Pujol, 1988). With regard to G. muellerae, peaks observed from 12.5 to 10.5 kyr are in agreement with Uk’37-estimated SST low values. However, and apart from the peak seen at 7.7 ka cal. BP, its NAR decreases from 10.5 ka cal. BP onward despite of the general cooling trend. From 7.7 ka cal. BP to the top, G. muellerae and small placoliths show similar variability. Small placoliths are eurithermal species linked to nutrient presence (Okada and Wells, 1997). Both taxa could have been controlled by the same factors since they show similar responses. Flores et al., (1997) found low abundances of G. muellerae after 73.9 ky in the Western Mediterranean, suggesting that the affinity of this species for cold-water conditions in the Alboran Sea is unclear. From our results, we suggest that abundance of G. muellerae has been responding to nutrients availability rather than SST variability since 7.7 ka cal. BP. With regard to WWG, it has to be mentioned that its signal results from the sum of all those taxa with preference for warm waters, something that not exclude that other environmental parameters such us salinity, nutrients, coastal currents, eddies and species-specific biogeography could affect their distribution and variability (Baumann et al., 2005) and thereby its signal in the paleorecords. Except for the period from 7.7 to 6.5 kyr, general trend of total NAR is in good accord with Uk’37-estimated SST record. This is due to total NAR is mainly constituted by NAR of small placoliths, including the alkenone-synthesizing coccolithophore E. huxleyi. From 7.7 to 6.5 kyr, total NAR records its highest values pointing to the highest productive period in the study area during the Holocene. The absence of agreement between this record and Uk’37-estimated SST record during this period suggests that factors other than SST controlled coccolithophorid production and preservation in deep-see sediments at that time. Except for G. muellerae, which shows high NAR values and a decreasing trend until 7.7 ka cal. BP, NAR of all taxa remain low and show a slight increasing trend. From that time to 7.7 to 6.5 ka cal. BP, all species record higher NAR values pointing to more productivity at that time. After 6.5 ka cal. BP, they show a slow decreasing trend up core punctuated by several peaks, suggesting a gradual reduction in productivity across the Holocene although affected by high variability. F. profunda and small G. oceanica show similar general variability from that time. F. profunda is abudant under a deep nutricline and an impoverished UPZ in nutrients (Molfino and McIntyre, 1990) something interpreted in terms of water column stability or stratification. F. profunda habis the LPZ, while G. oceanica is a warm-water species by many authors (e. g. Geitzenauer et al., 1976; Weaver and Pujol, 1988) living in the UPZ. Due to their ecological requirements, an opposite trend would be expected. Their paralellism is interpreted as a consecuence of pulses of water column stratification. This fact would promote enough nutrients for development of F. profunda under the thermocline. The scarcity of nutrients in the UPZ would prevent the flourishment of other species. Nevertheless, small G. oceanica would find the ideal conditions for its development above the thermocline in the warm NASW entering the Alboran Sea. Bбrcena et al., (2004) studied the annual flux of organisms measured by two deployed mooring lines on the continental slope south of pporting our hypothesis, these authors found that its flux decreased when wind-induced upwelling became more intense, relating G. oceanica to surface water stability and to the oceanographic evolution of the AJ. This relationship with the AJ was also suggested by Бlvarez et al., (2010) in a study of surface sediment samples from the Mediterranean Iberian Margin. In the same manner, Knappertsbusch (1993) and Cros (1995) related this species to superficial less saline Atlantic waters entering the Alboran Sea describing G. oceanica as a tracer of Atlantic surface waters in this region. Therefore, we interpret that the parallel record of small G. oceanica and F. profunda is due to the combined effect of the entrance of minor and warmer volume of flow of the AJ along with a stable water column.

5.3.  Short-term changes

5.3.1.  7.7 ka cal. BP event

At 7.5 ka cal. BP there is a simultaneous increase in NAR of all important species (e. g. small placoliths, F. profunda, Umbilicosphaera sibogae and G. muellerae). Because all taxa were affected, despite some of them have different ecological preferences (e. g. nutrients availability and SST) we argue that the same (or at least one) ecological factors control them. We propose that the possible cause of this peak in NAR is linked to the increase of the Atlantic inflow in the Western Mediterranean Sea and the ensuing establishment of WAG and EAG in the Alboran basin. Directly associated to the appearance of WAG, the productive cell ‘Malaga upwelling’ would be settled at that time, implying a nutrient input that would favour the blooming of all species, including F. profunda which, despite of living in the LPZ and stratified waters, could have taken advantage of the turbidity in the UPZ and appearing in conditions of low light in the LPZ. The proposed timing for the establishment of the WAG is in agreement with that proposed by Rohling et al., (1995), who reported this event around 8 ka BP from the interpretation of an abrupt faunal change in the planktic foraminiferal assemblage. Until 7.5 ka BP the global sea level was rising at a sustained rate (Bard et al., 1996) and the maximum flooding of southern coast of Spain is dated at 7.4 ka cal. BP (Zazo et al., 1994). Thereby, we suggest the latest forcing cause for the 7.7 ka cal. BP event is likely related to the latest remarkable fast-flowing AJ inflow triggered by the final melting of the remaining glacial ice in the North Atlantic. The increment seen in NAR of all coccolithophore species is synchronous to the faunal change identified in the planktonic foraminiferal assemblage at 7.7 ka cal. BP in the Alboran Sea (Pйrez-Folgado et al., 2003) and in the benthic foraminiferal assemblage recognised at 8 cal ka cal. BP. in the Gulf of Lions (Melki et al., 2009). A remarkable redox event in the Algero–Balearic basin deduced from Mn/Al ratio was identified at between 7.5 and 7.0 kyr cal. BP (Jimenez-Espejo et al., 2007). This redox event was interpreted by the authors to be originated by the redoxcline reaching the seabed due to intensification of the thermohaline circulation and/or the first moment in which WMDW fills its deep basins. Cortйs Sбnchez et al., (2012) reported climatic changes at that time from the review of many paleoenvironmental records from various sites in the basin and continental margin of the Alboran Sea. These changes were interpreted as a long-term environmental crisis due to sea-level rise and changes in the thermohaline circulation. Based on our results, we suggest that 7.7 ka cal. BP event observed in our records corresponds to the onset of the “Malaga upwelling” productive cell as a consequence of the establishment of WAG and its geostrophic front. This, ultimately, would be due to a major inflowing AJ entering the Alboran Sea at that time.

5.3.2.  Variability of the productive cell “Malaga upwelling”

High values of N ratio point to high productivity, while lower values indicate low productivity. Thereby, variability reported by N ratio (Figure 3) itself shows successions between stratification and upwelling pulses from 7.7 ka cal. BP across the Holocene. It has to be mentioned that N ratio shows high values across the whole period, changing just in a range of 10%. Nevertheless, Hernбndez-Almeida (2006) found similar N ratio values (minimum of 0.91) from analyses of sediment trap samples when satellite images and SST data pointed to strong thermal stratification in the Alboran Sea. From this, we deduce a stratified water column when N ratio decreases. Consequently, during N ratio peaks, upwelling pulses occurs. Vertical mixing is one of the most important proccesses accountable for the nutrient input in the study area (Dafner et al., 2003). In this sense, the deepening/shoalling of the nutricline, which is determined by the depth of MAW and Deep Waters (DW) interface (Echevarrнa et al., 2002), plays an important role. The depth of this interface is highly influenced by the rate of WMDW formation in the Gulf of Lions, which flows with LIW (both forming DW) inmediately under MAW. Frigola et al. (2007) analysed grain size and geochemical ratios for the last 12 ky from the marine core MD99-2343 located north of the island of Minorca (Figure. 1 c). From results of grain size distribution, like UP10 fraction (Figure 5a), the authors interpreted the occurrence of nine episodes, named as Minorca abrupt events (M events), and belonging to periods of WMDW reinforcement in the Gulf of Lions (Figure 5). A comparison between M events from the Gulf of Lions and N ratio record from the Alboran Sea (Figure 5c) shows a good correlation. The N ratio shows incresing trends during M8, M7, M6, M5 and M4. M3 and M2 do not show large increases in N ratio, but they both are separated by a large peak of low values, pointing to a strong stratification pulse that separates both events. M0 also registers a rise of N values, although it sudden records low values. From this, we interpret upwelling events in the study area whenever a M event occurs. The NAR of F. profunda —a species usually used in paleoceanographic reconstructions as an indicator of the nutricline depth and of the impoverishment in nutrients of the UPZ (Molfino and McIntyre, 1990; Beaufort et al., 1997)— has been also compared to M events (Figure 5d), showing good anticorrelation. Apart from M6, it shows low NAR values during every M event. Nevertheless, the N ratio shows an increasing trend during this event, showing a better expression of higher productivity during M6. F. profunda exhibits large peaks out of M events, specifically between M8-M7, M7-M6, M5-M4, M3-M2, M2-M1 and a smaller one between M4-M3, pointing to stratified water column off M events. Alcohol index has been previously related to sedimentary conditions in the Alboran basin (Cacho et al., 2000; Cacho et al., 2002). Since n-hexacosanol is more labile to degradation processes than n-nonacosane, decreases in the alcohol index can be interpreted as poor preservation of the organic matter in deep sea sediments. From 7.7 ka cal. BP, this index (Figure 5b) points to a general better preservation of the organic matter, likely due to a gradual decrease of deep water ventilation. This general trend is interrupted by several decreases, coincident with M5, M4, M3, M2 and M0 suggesting better deep water ventilation in the study area during these periods, something that could be interpreted as a consequence of the WMDW reinforcement that ventilates the seafloor off the coast of Malaga on its way to the Atlantic Ocean. However, this interpretation must be taken with caution since the Alcohol index has not been established yet as a robust deep water ventilation proxy in the study area. From the comparisons of our results with M events, we observe a strong in-phase link between long-lasting productive events in the Alboran Sea and WMDW reinforcement in the Gulf of Lions. We interpret that the more WMDW is formed in the Gulf of Lions, the more the interface rises (understood here in terms of nutricline) in the study area. Besides, and in order to compensate the major outflow at the Strait of Gibraltar which acts as a bottleneck, the AJ is strengthened. This along with a shoaling of nutricline would prompt vertical mixing, and thus, produce upwelling events. Consequently, a slackening of WMDW formation would lead to a deepening of the nutricline and a reduction of the AJ, encouraging stratification events. This fact supports our previous hypothesis, where a major/minor influx of AJ was deduced from the jointly absence/presence of small G. oceanica and F. profunda, respectively. Percentage abundance variability reported by reworked specimens (Figure 5e) exhibits large peaks during M8, M7, M6, M5, M4, smaller peaks during M3 and M0, and a large peak between M3 and M2. Reworked nannoliths have been previously linked in the Alboran Sea to terrigenous input belonging to exposed continental margins (Flores et al., 1997; Colmenero-Hidalgo et al., 2004). On the other hand, M events were suggested to occur when westerlies were stronger and climate conditions were dry (Frigola et al., 2007). Arid conditions during these periods are recognized in the Western Mediterranean from polen analysis of the marine core MD95-2043 by Fletcher et al., (2012) (Figure 5f). Decreasing abundance of Mediterranean and temperate pollen grains points to a decline in forest and more arid conditions during M1, M2, M3, M5 and M6. Despite dryer conditions imply minor precipitation over the region, we suggest fluvial discharge as the transport mechanism; arid conditions could lead to a reduction of the vegetation cover and hence, exposure of the continental margin, more easily erodible by the rivers flowing into the study area: Guadalfeo River and Guadalhorce River (Figure 1b). ­­The large peak seen in reworked nannoliths between M2 and M3 is interpreted as an increase in precipitation, leading to major river flow volume accountable for the terrigenous input at that time. This is supported by the large peak shown by pollen grains at that time (Fletcher et al., 2012) (Figure 5f), which suggests more humid conditions. The Mg/Ca-estimated SST record (Figure 5g) shows that the study area was affected by cooling trends during M8, M7, M5, M3, M2 and M1. However, M6 shows a peak in SST, M4 exhibits a clear warming trend and M0 shows stable values. Despite of M6 and M4, we interpret that drops in SST during M events are related to the long-lasting upwelling of colder waters in the upper water column.