An important component of 4R phosphorus fertilizer management is consideration of fertilizer source (form) and its interaction with the other 4R’s of rate and placement. Furthermore, localized landscape and soil characteristics have historically not been considered in prescription of 4R practices, yet are needed to more precisely manage P in variable landscapes. A three-year wheat (Triticum aestivum L.)- pea (Pisum sativum)- canola (Brassica napus L.) rotational study was initiated in 2021 across three field sites in Saskatchewan, Canada, with each site situated on a unique landform element (Upslope (knoll), Midslope (level), and Lowslope (depression). Agronomic and environmental performance was assessed as influenced by source: eight P fertilizer sources and an unfertilized control, rate: 20 vs 40 kg P₂O₅ ha^-1, and placement: side-band vs spring broadcast. Grain yields, P uptake and post-harvest residual soil available P were measured in a replicated experiment along with soluble P concentration in simulated snowmelt runoff. Grain, straw yield and P export in runoff varied across sites by P source, landscape position and soil properties. Less soluble sources like rock phosphate and struvite resulted in lower crop P uptake and residual soil P, but also showed lower P export in runoff across all sites. In both Upslope and Lowslope sites, polymer coated monoammonium phosphate produced the highest residual soil available P after three years, possibly reflecting reduced fixation due to slow release. Side-band placement was slightly advantageous over spring broadcast in producing greater crop P uptake and apparent P recovery and reduced total P load in runoff. Overall, Upslope and Midslope sites where yield and crop P removal over the crop rotation was lower and where surface soils became enriched with soil test P experienced greater losses of P in snowmelt runoff.