Welcome to the team, Dave 😀! I would start with the fence in the sunlight at right:

It’s relative close and therefore detailed and it shows at least one complete fence element. I think, the fence can help later forming the landscape by looking at the angles between the fence and the ground. I would first model these elements so that they roughly match the shape in the artwork and then align with the camera’s perspective. With the house blocked out by Kent and aligned with the artwork you can orbit around the objects and check whether their relative positions in the 3D scene make sense. When I blocked out the house last month, I often had the problem that objects looked well aligned with the artwork from the camera’s perspective but bad when seen from a different angle.