Abstract. Over the 20th century, surface water temperatures have increased in many lake ecosystems around the world, but long-term trends in the vertical thermal structure of lakes remain unclear, despite the strong control that thermal stratification exerts on the biological response of lakes to climate change. Here we used both neo-and paleoecological approaches to develop a fossil-based inference model for lake mixing depths and thereby refine understanding of lake thermal structure change. We focused on three common planktonic diatom taxa, the distributions of which previous research suggests might be affected by mixing depth. Comparative lake surveys and growth rate experiments revealed that these species respond to lake thermal structure when nitrogen is sufficient, with species optima ranging from shallower to deeper mixing depths. The diatom-based mixing depth model was applied to sedimentary diatom profiles extending back to 1750 AD in two lakes with moderate nitrate concentrations but differing climate settings. Thermal reconstructions were consistent with expected changes, with shallower mixing depths inferred for an alpine lake where treeline has advanced, and deeper mixing depths inferred for a boreal lake where wind strength has increased. The inference model developed here provides a new tool to expand and refine understanding of climate-induced changes in lake ecosystems.

Accurate estimates of Stokes drift are necessary to quantify Lagrangian transport and upper-ocean mixing. These can be estimated from directional wave spectra. Here, a methodology for the reconstruction of such spectra is developed using partitioned bulk wave parameters provided by global wave models. These reconstructed spectra agree well with global wave model–simulated full spectra. Regional wave model simulations with reconstructed spectra as open boundary conditions lead to more accurate estimates of bulk wave parameters in the coastal ocean. Furthermore, the reconstructed directional spectra can be used to improve high-frequency (HF) radar–derived surface Lagrangian current estimates. Stokes drift vertical profiles from complete directional spectra are more accurate, and therefore coupled ocean circulation and wave models should incorporate spectral estimates for wave–current interaction studies. Based on model simulations conducted here, it is recommended that regional wave modeling studies use partitioned rather than bulk wave parameter products from global wave simulations to reconstruct complete directional spectra for open boundary conditions. Finally, this study shows that inclusion of the peak spectral energy for each partition improves the ability to reconstruct more accurately directional spectra and surface Stokes drift. It is recommended that the global wave model hindcast/forecast include this additional bulk parameter.

No abstract

The formation of coastal dense shelf water in winter provides the available potential energy (APE) to fuel baroclinic instability. The combined effects of baroclinic instability and wind forcing in driving cross-shelf exchange are investigated using idealized numerical simulations with varied bottom slope, wind stress, and heat loss rate. The results show that under upwelling-favorable winds, the intensity of the instability decreases as the wind stress increases. This is caused primarily by enhanced turbulence frictional dissipation. Under downwelling-favorable winds, an increase in wind stress and/or a decrease in heat loss rate tends to constrain the baroclinic instability, leading to a circulation resembling that driven purely by wind forcing. In the latter case, once a critical value of cross-shore density gradient is reached, isopycnal slumping is initiated, leading to increased vertical stratification and narrowing of the inner shelf. The change in depth of the inner-shelf outer boundary, defined as the location corresponding to the maximum cross-shore gradient of the surface Ekman transport, is proportional to an empirically derived multiparametric quantity , where a2 is a dimensional constant, B0 is a constant heat loss rate, γ = 0.43, f is the Coriolis parameter, α is the shelf slope, B is the heat loss rate, and τ is the wind stress. This relationship is found to hold for cases when instabilities are present.

A hybrid, empirical radar wave inversion technique that treats swell and wind waves separately is presented and evaluated using a single 48-MHz radar unit and in situ wave measurements. This hybrid approach greatly reduces errors in radar wave inversion during swell seas. Our analysis suggests that, prior to the inversion, the second-order spectrum should be normalized using Barrick’s weighting function because this process removes harmonic and corner reflection peaks from the inversion and improves the results. In addition, the resulting calibration constants for the wind wave component are not wave-frequency dependent and are similar in magnitude to those found in previous studies using different operating-frequency radars. This result suggests radar frequency independence, although additional experimental verification is required. The swell component of the model presented here ignores the effect of swell’s propagation direction on the radar signal. Although this approach has several limitations and may only be useful near the coast (where swell propagates close to perpendicular to the coastline), the resulting wave inversion is accurate even when swell is propagating close to perpendicular to the radar beam direction. RMS differences relative to in situ wave height measurements range from 0.16 to 0.25 m as the radar beam angle increases from 22° to 56°.