Our goal is to predict interactions that develop among nephrons and between nephrons and the arterial network that supports them. We have developed a computationally simple but physiologically-based mathematical model of the kidney vascular tree to study renal autoregulation in ensembles of interacting nephrons not directly available for experimentation. The study combines computed tomography (CT) of a renal vascular cast at 2 micrometer resolution with simulation. The CT scan showed a bifurcating branching structure with as many as 7 bifurcations between arcuate arteries and the renal surface, with afferent arterioles originating from all arterial structures, including arcuate arteries. The modeling component has 2 novel features: a probability based vascular tree based on the data from the CT images, and a network of arteries supplying several simple whole nephron models coupled electrotonically. The network model predicts dynamical aspects of vascular pressure drops and nephron self-sustained cooperative dynamics.