Large-scale entry of intermittent renewables needs to be balanced with a source of flexibility to retain generation adequacy. Storage offers such flexibility but sufficient capacities are currently available only indirectly via hydropower reservoirs. Yet hydropower is not insulated from adequacy concerns because reservoir levels also depend on the weather. This paper shows how to estimate the availability of intermittent renewables and long-term hydro storage from historic data with the use of modern regression techniques. The novelty of the paper is to introduce a robust methodology to add the estimated availability distributions directly to the computationally light and transparent recursive convolution approach; and the principles extend to other computational methods. An application to the Finnish power system makes it clear how both the short-term availability shocks from intermittent renewables and the long-term shocks from hydropower availability should be properly accounted for to avoid misguided generation adequacy estimates.