Projected future biophysical states of the Bering Sea

Abstract

Three global climate simulations from the Intergovernmental Panel on Climate Change Fourth Assessment (AR4) were used as physical forcing to drive a regional model that includes both physical and biological elements of the Bering Sea. Although each downscaled projection indicates a warming of 1–2 °C between 2010 and 2040 on the Bering Sea shelf, the interannual and interdecadal details of this trend vary considerably among the three realizations. In each case, the magnitude of presently observed interannual variability of bottom temperatures and ice cover is found in the models to be maintained out to at least 2040, but with a steadily increasing probability of warm years with less ice on the southern shelf. The overall trends indicate warmer temperatures and the retreat of ice in the southeastern Bering Sea, but continued ice cover in the northeastern Bering Sea. Sensitivity analyses suggest both increasing air temperature and northward wind stress as primary drivers of higher water-column temperatures. Based on currently available models, changes in shortwave radiation are not likely to have a significant role in this warming. Warming trends on the outer shelf may lead to decreased production of large crustacean zooplankton at that location, but could increase such production on the inner shelf.

Introduction

Prominent physical features of the Bering Sea include seasonal ice cover, strong advection of ice, and tidally generated biophysical domains. Ice formed each winter in the northern Bering Sea is advected to the southeast, where it gradually melts as it encounters warmer water and air temperatures. This southward advection contributes to the latitudinal salinity gradient of the Bering Sea and its interannual variability. A cross-shelf gradient in the vertical penetration of tidal mixing sets up distinct biophysical regimes with associated biota. Classically, the southeastern shelf is classified as having three biophysical domains: a vertically well-mixed inner region (~0–50 m), a middle region which is well-mixed in the winter and has two distinct layers separated by a sharp thermocline in the summer (~50–100 m), and an outer region which is more gradually stratified (~100–200 m) (Kinder and Schumacher, 1981, Coachman, 1986, Kachel et al., 2002).

Distinct biological features of the Bering Sea ecosystem include ice algae as a potential food source to secondary producers, and strong benthic–pelagic coupling. Within the different biophysical regimes, the relative magnitude of pelagic vs. benthic pathways of carbon flux varies interannually, and is believed to be strongly influenced by the extent of seasonal ice through its effects on stratification (Hunt et al., 2002, Hunt et al., 2011). The relative importance of pelagic vs. benthic pathways is likely to shift under the influence of global warming, partially through its impact on seasonal ice extent in the Bering Sea. Field data suggest that recent cold temperatures in the Bering Sea have led to an increase in large crustacean zooplankton, favored as food items by juvenile pollock in the fall season (Coyle et al., 2011).

The present hydrography and climatology of the Bering Sea result in a highly productive ecosystem, with plankton biomass ultimately supporting large populations of shellfish and finfish (and major fisheries), marine birds and marine mammals. Such intense production derives, in part, from a broad shelf with strong tidally induced mixing, a plentiful supply of the micro-nutrient iron, and seasonal stratification which maintains the phytoplankton in the euphotic zone, adjacent to a deep, macronutrient-rich basin. Cooling trends in the Bering Sea from 2006 to 2011 (Stabeno et al., 2012a) have been documented by the Bering Sea Ecosystem Program (BEST), the Bering Sea Integrated Ecosystem Research Program (BSIERP), the U.S Bering-Aleutian Salmon International Survey (BASIS), and the North Pacific Climate Regimes and Ecosystem Productivity Program (NPCREP). Measurements since late 2013 indicate a return to warmer conditions, with reduced ice (Stabeno et al., 2016). The response of Bering Sea production to changes in temperature is not yet completely clear and may depend on the timescale under consideration. While ocean color observations over the Bering Sea suggest that primary production during warm years may be enhanced by 40–50% compared to cold years (Brown and Arrigo, 2013) it has also been suggested that very warm temperatures suppress summer production, because intense water column stratification (Coyle et al., 2008) reduces the re-supply of nutrients to the upper mixed layer.

A model-based multivariate analysis was used to help explore the relationships between physical and biological factors on the Bering Sea shelf (Hermann et al., 2013, henceforth referred to as H2013). The analysis suggested that the Bering Sea shelf may not respond uniformly to changes in climate forcing. For example, large crustacean zooplankton (LCZ) are negatively correlated with temperature on the outer, southwestern shelf, and positively correlated to temperature on the inner, northeastern shelf. Areas of positive correlation tend to correspond with those areas with greatest change in ice cover. As in the revised Oscillating Control Hypothesis of Hunt et al. (2011), the ratio of large to total zooplankton is enhanced at lower temperatures. On the outer shelf, higher temperatures may be leading to reduced LCZ production either through effects on stratification (and hence nutrient limitation), or through direct effects of temperature on growth, respiration, predation and vertical migration. Changes on the northern shelf may involve a complex interplay of light and nutrient limitation effects, as modulated by a reduction in the duration of seasonal ice cover.

Ice dynamics of the Bering Sea have been explored in both observational and modeling studies (Stabeno et al., 2010; Danielson et al., 2011a, Danielson et al., 2011b, Cheng et al., 2014; Li et al., 2014a, Li et al., 2014b; Sullivan et al., 2014). Ice is formed seasonally in the northern Bering Sea and is advected southward, resulting in a net transfer of freshwater from north to south. Heat budgets from these studies have underscored the importance of sensible heat flux between the atmosphere and the ice in the northern Bering, and between the ocean surface and the ice in the southern Bering, where the ice edge retreats each spring.

We begin with a description of the global and regional models used for the biophysical modeling of this region, and the physical data used for comparison with the models. This is followed by a comparison of the model output with some of the moored and gridded data collected during 1971–2012. We next consider our three downscaled projections of future conditions in the Bering Sea, and highlight the most significant changes from present conditions. Finally, we consider which elements of the physical forcing appear most likely to govern the projected changes, using the coherence among relevant pairs of physical and biological features.