Depth-dependent soil phosphorus alteration is independent of 145-year phosphorus balances

Agricultural management practices can profoundly influence soil phosphorus (P), with effects accumulating over time. To test the overarching hypothesis that soil P pools estimated by sequential fractionation would be altered by long-term agricultural practices, we used an experiment established in 1876 in the north-central US to quantify 145-year impacts of crop rotation (continuous maize [Zea mays L.], maize-soybean [Glycine max L. Merr.] and maize-oat [Avena sativa L.]-alfalfa [Medicago sativa L.]) and 117-year impacts of fertilization (unfertilized and fertilized) with rock phosphate, manure or synthetic fertilizer on soil P fractions at 15 cm intervals across 0–90 cm depth. Fertilization impacts on soil P were mostly limited to the surface (0–30 cm) depth, but extended to 90 cm depth under diverse rotations. Under fertilization, soil total P concentration increased by 51% at 0-30 cm while concomitantly decreasing by 30% at 60–90 cm compared to no fertilization, indicating that vertically stratified surface soil P accumulation and subsoil P depletion can co-occur even under positive P balances. Positive P balances (1222–1494 kg/ha) induced by fertilization enriched inorganic P (P_i) (+39% to 358%) and labile organic P (P_o) fractions (+11%) while depleting non-labile P_o fractions (−31%), with depletion increasing with the degree of crop diversification. Fertilization had minor impacts on P fractions beyond 30 cm depth, except for acid extractable P_i (HCl-P_i) depletion under continuous maize and maize-soybean rotations (−16% to −78%) and accumulation under maize-oat-alfalfa rotation (+41% to +84%) at 60–90 cm. In contrast, without fertilization, diversifying maize rotations with oat and alfalfa decreased HCl-P_i and residual P (−21% to −57%) but increased non-labile P_o fractions (+54%), suggesting potential mining of non-labile P_i pools by deep-rooted legumes under nutrient limitation. The 1–2 orders of magnitude greater changes in stocks of P fractions than stocks of total P emphasize the importance of distinguishing P pools even with operational fractionation to fully capture changes in P cycling beyond total P stocks. Our study revealed that a positive P balance under 117 years of fertilization (i) enriched P_i and labile P_o pools but (ii) depleted non-labile P_o pools, (iii) largely at 0–30 cm, and (iv) non-labile P_o depletion increased with crop diversification under 145-year rotation treatments.