I think create a dam on a peatland is not easy because the terrain is very flat. What is the goal to measure runoff? Water quality, runoff peak? In some hydrologically altered wetlands a dam was built in order to increase the water-level but the effects on nutrient biogeochemistry becomes disappointing.
If what you need is to measure discharge, then you should be looking to install a weir with a v-notch. As to where install it, you'll need to identify the discharge point, usually an outflowing stream. V-notch are calibrated depending on the angle and from height on the crest one can easily calculate flow.
Are evapotranspiration and rainfall included in you study/model? Then, measuring water-level variation (using a pressure transducer such as level logger) you can calculate runoff as the unknown variable. Of course, it cannot separate among sources (surface or groundwater). A weir with a v-notch is usually hard to install in a wetland.
That sounds really good! As far as I know, data from continous (authomatic) water level monitoring incluided in a DEM is the only method to be able to have a continuous estimates of runoff. The temporal scale (hourly to daily) of this monitoting depends on how fluctuating the hydrology. As you pointed out using the word "catchment" and since you have several rain and lysimieter stations I infer that the extension of the peatland is large. That's right? It is important if streamflow is the main water source or when spatial distribution of rainfall is rather heterogeneous. Finally, estimates of evapotranspiration should also be included in order to close the water balance. I think it is important, mainly during the growing season.