Implicit cranck nicolson

Dear All,

I had written a cod on 1-D diffusion equation (parabolic PDE, with time and space) based on finite difference codewith implicit crancknicolson formulation. The problem is for solving rainfall induced soil-slope stability with intia and boundary conditions. I am comparing my results with already published work but, for some time of 6hr duration results are matching to some extent. When the time simulation exceeds 6hr the difference in result was obsered was more than 20 %. I had checked the formulation, intial and boundary conditions  it looks to be okay. I had maintained consistent units.

Please suggest me where am I wrong. Is there any stability and convergence issues with crankniconson implicit method (Central difference). I had even maintained timestep to spacestep square ration almost equal to 1.

Please suggest me on any stability issues with CN method or any method which can give precise result or is there anything I can change in my code.

Thank you in advance.