Phosphorus losses from agricultural land to surface waters : impact of grazing and trampling by horses

Abstract: The number of horses in Sweden has increased in recent decades, reaching over 300,000 in 2009. The horses are kept on 300,000 ha representing 10% of total agricultural land in Sweden. This study characterised the potential risk of phosphorus (P) losses from a heavy clay soil used for horse grazing and feeding (paddock) and compared the losses with nearby arable land managed conventionally and losses from ungrazed pasture. Water-soluble phosphorus (WSP) concentration in surface soil (0-10 cm) from the paddock areas (mean 0.62 mg 100 g-1 soil) did not differ significantly from that in arable land, but differed very significantly (p<0.001) from reference areas with ungrazed pasture. Phosphorus extracted in acid lactate (P-AL value) in the topsoil of paddock (mean 14.7 mg 100 g-1 soil) differed significantly both from arable land and ungrazed pasture (p=0.031 and 0.033, respectively). Total phosphorus extracted with nitric acid (P-HNO3) in topsoil of paddock areas (mean 117.3 mg 100 g-1 soil) did not differ significantly from arable land (p=0.08) but was significantly higher (p<0.001) than in ungrazed pasture. Paddock soils with high levels of different P forms, thus, pose a high risk of P losses. Aluminium and iron content in acid extract (Al-AL and Fe-AL) were found to be highly correlated to P-AL (r=0.76, p<0.001; r=0.75, p<0.001, respectively), indicating no significant difference for the dominance of Al over Fe for P availability in the fields studied. No good correlation was found with Ca (r=0.26, p=0.03) but soil C content was found to be correlated with P-AL (r=0.46, p<0.001). In the past 8 years, high P concentrations (up to 1.5 mg L-1), mainly in dissolved reactive form, have been recorded in drainage water from the catchment (30 ha). Therefore, we concluded that horse grazing at high stocking rates (>2.5 livestock units ha-1) may pose a risk of high P losses to nearby water bodies.