Summer drought is a feature of the semi-arid region of central Oregon, USA, where vegetation naturally develops into ponderosa pine (Pinus ponderosa var. Laws) forest. Forest management consists of clearcut harvest and natural regeneration. Soil microbial activity is interconnected with forest processes because substrate quality and availability can be important driving variables. Stand development influences the soil water regime, and water availability may also limit microbial activity. We determined factors limiting litter and mineral soil carbon (C) mineralisation rates in undisturbed old growth and regenerating (hereafter, young) ponderosa pine stands under a semi-arid climate. Mass of litter and dead fine roots did not differ significantly between the stands, but litter substrate quality was different. Young stand litter had significantly higher concentrations of total nitrogen (N), extractable organic N, extractable C, and microbial C and N than that from the old stand, probably because of litter fall from the broadleaved shrub understorey, including the N-fixing species Purshia tridentata (Pursch) DC, that comprised 40% of the young stand’s leaf area. The old stand contained no understorey. For litter samples from the two stands, wetted to 60% of water-holding capacity (WHC), net mineral-N and CO2–C mineralisation rates were similar despite the substrate quality differences. Mineral soil properties at 0–0.1 m depth were similar in the two stands, except for lower CO2–C production in samples from the young stand; at 0.1–0.5 m depth, total C and N and microbial N concentrations were higher in the young stand. Net mineral-N production in field-moist soil, sampled during a typical summer drought and incubated at 25 °C for 56 days, was generally 3–6 mg kg−1 soil at both sites, but increased up to 29 mg kg−1 upon wetting to 60% of water-holding capacity. Over 56-day-long incubations, wetting also increased litter and soil microbial respiration rates by factors of about 500 and 3, respectively. The incubations yielded a proportionality between respiration rate and water content that was supported by in situ measurements of soil respiration in the young stand, before and after irrigation. A hypothetically wet year without soil water deficit caused a 2.5-fold increase in a modelled estimate of the young stand’s annual soil respiration rate. Litter and soil C mineralisation rates in these ponderosa pine forests thus appeared to be limited much more by the availability of water than by a lack of available C or N substrates.