Phosphorus (P) availability to soil biota is commonly determined using soil extraction methods, which can mobilise unclearly specified and stable soil P fractions. Alternatively, natural P availability can be estimated using in situ exposition of ferrous-oxide impregnated resin that continually removes P from soil solution. Over 10 consecutive years, we measured phosphate dynamics using such resin in soils of two catchments, differing in their bedrocks (P-poor mica-schist and P-rich granite), P-sorption characteristics (different Al- and Fe-hydroxide pools) and terrestrial P export to receiving waters. The catchment with granitic bedrock loses P in the long-term, while soils on mica-schist permanently accumulate P from atmospheric deposition. We observed substantially higher soil P availability in the organo-mineral horizons along with higher terrestrial P export in the granitic catchment rather than in the mica-schist one. These results indicate that soils developed on the P-rich bedrock, and concurrently with lower P sorption capacity, have high in situ P availability. We further evaluated the main factors affecting soil P availability, using data on long-term environmental and edaphic variables. Soil P availability was partly controlled by abiotic factors such as throughfall chemistry, precipitation amount and the C-to-P ratio of litter. However, we assume that high losses of P in the granitic catchment can also be driven by microbial processes since decomposition activity was the most influential variable of available soil P in this catchment.