Canary current upwelling: More or less?

Highlights

• Evidence for a reported recent increase of coastal upwelling off NW Africa is examined.

• No significant increase of upwelling-favorable winds is taking place.

• Sea surface temperatures records indicate warming, not cooling, of the coastal zone.

• The apparent discrepancy in the alkenone-temperature relation is explained.

• No intensification of the upwelling is found.

Abstract

It has been hypothesized that coastal upwelling in the four major eastern boundary current systems might be intensified as global warming could result in a greater land-sea temperature gradient and hence strengthen alongshore winds. Recent research has suggested a substantial increase of upwelling intensity off Northwest Africa. Evidence there is based on the derivation of a proxy for upper ocean temperatures from the alkenone unsaturation index (${\text{U}}_{37}^{{\text{K}}^{\prime }}$) derived from two sediment cores recovered off Cape Ghir, Morocco. An accelerating decrease of over 1 °C during the last century was concluded for near surface temperature near the Cape. Support for this conclusion was found in an increase in Bakun’s upwelling index for the same area. The evidence for a general intensification of upwelling within the whole Canary current upwelling system is examined here. Using available estimates of wind from PFEL, NCAR/NCEP, ECMWF, ICOADS and WASWind plus measured wind data from coastal meteorological stations, no evidence of a coherent intensification in winds at the regional scale off Northwest Africa is found. Moreover, sea surface temperature records from ships-of-opportunity (ICOADS data set) and also from the Pathfinder satellite AVHRR data set show a significant and correlated increase at all latitudes in the region, including in the area around Cape Ghir. It is concluded that there is no evidence for a general increase in upwelling intensity off Northwest Africa or Iberia. The apparent lowering of SST off Cape Ghir indicated by the alkenone unsaturation index can be explained by coccolithophorids (phytoplankton from which the ${\text{U}}_{37}^{{\text{K}}^{\prime }}$ signal is derived) living deeper in the water column. The distribution of most phytoplankton (including coccolithophorids) will deepen and have less overlap with mixed layer temperatures as the ocean warms, resulting in a near surface temperature estimate that is increasingly biased by subsurface temperatures and lower than the actual SST.

Introduction

An apparent paradox associated with the idea of global warming is that it might lead to intensified cooling of local areas affected by coastal upwelling (Bakun, 1990). This latter phenomenon takes place along the eastern boundaries of temperate and subtropical oceans where winds predominantly blow equatorward and alongshore. The combined effect of the wind stress and the rotation of the earth produces a net offshore transport of water in the upper layer of the ocean that must be replaced by a vertical motion of colder subsurface water near the coast – upwelling (Smith, 1968). Upwelling is important for renewing the nutrient content of the upper photic layers and supports some of the most productive regions of the oceans (Ryther, 1969). Indeed, some 20% of global marine fisheries catch originates in these areas, which represent only 5% of its total extent. However, upwelling regions are highly variable and even understanding the sign of change in these regions (either warming or cooling) will be a major contribution to predicting future ocean change and fisheries productivity in these regions.

Bakun (1990) outlined a scenario wherein increased atmospheric CO₂ content would diminish nocturnal cooling and enhance diurnal warming of the continental land masses at the latitudes of the upwelling regions. The strengthened thermal low pressure areas would enhance the net atmospheric pressure difference between land and ocean in these regions, thus producing an intensification of the equatorward wind and therefore of upwelling in the coastal ocean. Bakun (1990) presented evidence of a tendency for stronger upwelling in the Californian, Iberian, Moroccan and Peruvian areas during the period from 1945 to 1985. His analysis was based on the coastal upwelling index, or rate of upwelling per unit length of coastline, which is calculated from the geostrophic wind estimated from analyzed surface atmosphere pressure fields. The index has been in wide use as an indicator of upwelling strength ever since its introduction by Bakun (1973).

More recently, McGregor et al. (2007) found strong support for the hypothesis through evidence of a dramatic lowering of inferred SSTs in the northwest African upwelling region over the last century. From analysis of two cores taken some 25 km off Cape Ghir, Morocco, they derive a proxy record of ocean temperature, considered to reflect SST, over the last two and a half millenia. Their alkenone unsaturation index indicates a reduction of sea surface temperature (SST) in the Cape Ghir upwelling of 1.2 °C over the last century, coincident with the observed increase in atmospheric CO₂ levels. Wind observations from the COADS – Comprehensive Ocean–Atmosphere Data Set Release 1 were used to show an increase in wind over the zone in the period after 1950. Further supporting evidence was provided by the calculated upwelling (Bakun) index for the area over the same period.

However, the evidence that wind velocity is increasing and SST decreasing in the NW African upwelling region as a whole has many uncertainties. For example, the reported northward encroachment of more sub-tropical and tropical species of fish (Quéro et al., 1998, Brander et al., 2003, Graham and Harrod, 2009) is incongruent with decreasing temperatures in this coastal zone. In global studies of trends in Large Marine Ecosystems, including the Canary Current LME (Belkin, 2009, Sherman et al., 2009), it was concluded that the “most striking result is the consistent warming” by amounts exceeding 0.6 °C over 25 years in mid to low latitudes. Other more detailed studies like Narayan et al. (2010) and Pardo et al. (2011) have examined long term variation in the Canaries and other upwelling zones with varied results. Pardo et al. found “weakening of the upwelling intensity in the Iberian/Canary and NW African regions”. In contrast, Narayan et al. concluded that coastal upwelling intensity was increasing based on evidence for increasing upwelling-favorable winds and an upwelling index based on the difference between offshore and nearshore SSTs.

As different analyses give conflicting results, the publicly available wind and SST data sets for the North Atlantic upwelling region (Fig. 1) between Dakar, Senegal (15°N) and Galicia, Spain (43°N) are examined to document temporal trends in the intensity of the Canary Current upwelling regime in terms of the atmospheric forcing and the oceanic response. Our results clearly indicate that there is no evidence for increasing intensity of upwelling along this seaboard, in direct contrast to earlier studies. The many factors affecting reliability of the different sources of wind and SST time series are considered, before going on to explore the nature of the SST-alkenone proxy temperature relationship. The difference between the proxy and directly observed SST can be reconciled by consideration of the response to increasing stratification of the phytoplankton from which the ${\text{U}}_{37}^{{\text{K}}^{\prime }}$ signal is derived.