Converges at second order (up to 256^2). Errors in q flatten out at higher resolutions at ~10^(-10).
Finite differencing artifact since we require the connection coefficients to compute the Eckart heat flux?
Is this the best we can do? Convergence plot in grim considers 512^2 but grim shows larger base errors, due to which the L1 norm at 512^2 ~10^(-8)
Converges at second order (up to 256^2). Errors in q flatten out at higher resolutions at ~10^(-10). Finite differencing artifact since we require the connection coefficients to compute the Eckart heat flux?
Is this the best we can do? Convergence plot in grim considers 512^2 but grim shows larger base errors, due to which the L1 norm at 512^2 ~10^(-8)