Multi-plane PIV measurements were performed in an open-channel flume filled with elongated prisms of height k and width to investigate the effect of the deepening of the canopy on the flow structure. Velocity measurements were performed both inside the canopy and above it. Analysis of the spatial convergence for the double-averaged quantities shows that for canopy flow investigations (z < k), at least 5 measurement planes are required to obtain a relative spatial convergence error below 3% for the dispersive shear stress, the quantity the most sensible to spatial sampling. With only three measurement planes, the spatial convergence is below 1% only in the flow region above the canopy (z > k). Three canopy aspect ratios, k/ℓ = [1, 3, 6] were investigated for a fixed modified-submergence ratio β = (h − k)/ℓ = 3 where h is the water depth. As the canopy deepens, the hydraulic roughness decreases and the velocity near the bottom of the canopy becomes gradually constant, as expected for deep canopies. We show how the highly converged (both in space and time) profiles of double-averaged longitudinal velocity and total shear stress ... mehr can be used to calculate the vertical distribution of drag in the canopy. With this methodology, values of the drag coefficient CD(z) can be calculated, and are found to be always close to unity, even in the upper part of the canopy.