Helioseismology has revealed that the angular velocity of the Sun increases with depth in the outermost 35 Mm of the Sun. Recently, we have shown that the logarithmic radial gradient ($\rm d\ln\Omega/\rm d\ln r $) in the upper 10~Mm is close to $-1$ from the equator to $60^\circ$ latitude.We aim to measure the temporal variation of the rotational shear over solar cycle 23 and the rising phase of cycle 24 (1996-2015). We used f mode frequency splitting data spanning 1996 to 2011 from the Michelson Doppler Imager (MDI) and 2010 to 2015 from the Helioseismic Magnetic Imager (HMI). In a first for such studies, the f mode frequency splitting data were obtained from 360-day time series. We used the same method as in our previous work for measuring $\rm d\ln\Omega/d\ln r $ from the equator to $80^\circ$ latitude in the outer 13~Mm of the Sun. Then, we calculated the variation of the gradient at annual cadence relative to the average over 1996 to 2015. We found the rotational shear at low latitudes ($0^\circ$ to $30^\circ$) to vary in-phase with the solar activity, varying by $\sim \pm 10$\% over the period 1996 to 2015. At high latitudes ($60^\circ$ to $80^\circ$), we found rotational shear to vary in anti-phase with the solar activity. By comparing the radial gradient obtained from the splittings of the 360-day and the corresponding 72-day time series of HMI and MDI data, we suggest that the splittings obtained from the 72-day HMI time series suffer from systematic errors. We provide a quantitative measurement of the temporal variation of the outer part of the near surface shear layer which may provide useful constraints on dynamo models and differential rotation theory.

Helioseismology has provided unprecedented information about the internal rotation of the Sun. One of the important achievements was the discovery of two radial shear layers: one near the bottom of the convection zone (the tachocline) and one near the surface. These shear layers may be important ingredients for explaining the magnetic cycle of the Sun. We measure the logarithmic radial gradient of the rotation rate ($\rm d\ln\Omega/\rm d\ln r$) near the surface of the Sun using 15 years of f mode rotational frequency splittings from the Michelson Doppler Imager (MDI) and four years of data from the Helioseismic and Magnetic Imager (HMI). We model the angular velocity of the Sun in the upper $\sim 10$ Mm as changing linearly with depth and use a multiplicative optimally localized averaging inversion to infer the gradient of the rotation rate as a function of latitude. Both the MDI and HMI data show that $\rm d\ln\Omega/\rm d\ln r$ is close to $-1$ from the equator to 60$^{\circ}$ latitude and stays negative up to 75$^{\circ}$ latitude. However, the value of the gradient is different for MDI and HMI for latitudes above $60^{\circ}$. Additionally, there is a significant difference between the value of $d\ln\Omega/d\ln r$ using an older and recently reprocessed MDI data for latitudes above $30^\circ$. We could reliably infer the value of $\rm d\ln\Omega/\rm d\ln r$ up to 60$^{\circ}$, but not above this latitude, which will hopefully constrain theories of the near-surface shear layer and dynamo. Furthermore, the recently reprocessed MDI splitting data are more reliable than the older versions which contained clear systematic errors in the high degree f modes.

Studies of solar and stellar convection often employ simple polytropic setups using the diffusion approximation instead of solving the proper radiative transfer equation. This allows one to control separately the polytropic index of the hydrostatic reference solution, the temperature contrast between top and bottom, and the Rayleigh and Peclet numbers. We extend such studies by including radiative transfer in the gray approximation using a Kramers-like opacity with freely adjustable coefficients. We study the properties of such models and compare them with results from the diffusion approximation. We use the Pencil Code, which is a high-order finite difference code where radiation is treated using the method of long characteristics. The source function is given by the Planck function. The opacity is written as kappa=kappa_0 rho^a T^b, where b is varied from -3.5 to +5, and kappa_0 is varied by four orders of magnitude. We consider sets of one dimensional models and perform a comparison with the diffusion approximation. Except for the case where b=5, we find one-dimensional hydrostatic equilibria with a nearly polytropic stratification and a polytropic index close to n=(3-b)/(1+a), covering both convectively stable (n>3/2) and unstable (n<3/2) cases. For b=3 and a=-1, the value of n is undefined a priori and the actual value of $n$ depends then on the depth of the domain. For large values of $\kappa_0$, the thermal adjustment time becomes long, the Peclet and Rayleigh numbers become large, and the temperature contrast increases and is thus no longer an independent input parameter, unless the Stefan Boltzmann constant is considered adjustable. Proper radiative transfer with Kramers-like opacities provides a useful tool for studying stratified layers with a radiative surface in ways that are more physical than what is possible with polytropic models using the diffusion approximation.