I would like to add a cross-gradient term in the joint inversion and need to calculate the spatial difference of the model, but I am not so familiar with the operators. Specifically, for a 2D model, the tensor mesh contains cellGradx and cellGrady to calculate the cell centered gradient, but it takes us to cell faces. What if I need to compute the spatial difference? For example, for 1D vector x=[0,1,2,3,4], the spatial difference could be D.dot(x)=[1,1,1,1] if the dimension of D is 4x5. But I think typically, we will define D as 5x5 so we will still get a 5x1 vector. I didn’t find a differential operator that can achieve this. Could you please help to figure it out? Thanks.