Fate of phosphorus in soil during a long term fertilization experiment in southern Finland

The aim of this thesis was to study the accumulation and depletion of inorganic soil phosphorus (P) in a long term (37 yrs) field experiment in Southern Finland. The soil has a loamy texture and a high pH (7.4) due to heavy liming before the onset of the experiment. The field contains three treatments with different rates of P fertilization 0, 32 and 67 Kg ha-1 a-1 (abbreviated as P0, P1 and P2K) and it has been cultivated with a rotation of barley, oats, spring wheat and rye grass. The grain and straw have been collected every year and analysed for yield and mineral element composition. Soil samples were also collected from the plough layer at the beginning of the experiment in 1978, and also in the years 1995, 2005 and 2015, including also subsoil samples (25-50 cm). The changes in soil inorganic P along the years were studied with the Chang and Jackson fractionation The highest rate of P fertilization (P2K) increased the P concentration extracted with NH4Cl, NH4F, NaOH and H2SO4. These extractions are assumed to dissolve, easily available P, P sorbed by aluminium and iron oxides and calcium phosphates, respectively. On the other hand, the continuous P depletion (P0 treatment) decreased the NH4Cl-P and NH4F-P pools, while the NaOH-P and H2SO4-P pools remained stable. None of the P pools changed significantly at the lower rate of P fertilization (P1). The results indicate that the iron and aluminium oxides have a similar capacity for P sorption. However, P sorbed by iron oxides appears to remain non-available. Moreover, this study shows that in heavily limed soils the formation of secondary calcium phosphate compounds can be an important sink of the applied P. After 37 years, the P2K and P1K treatment increased the inorganic P reserves of the soil by 732 Kg P ha-1and 32 Kg P ha -1, respectively, while the P0 treatment depleted -459 Kg P ha -1. The recovery rates of P in the plough layer (i.e share of the P balance that is accumulated or depleted in the soil) were 79%, 30% and 56% for the P0, P1 and P2K treatments, respectively. An estimation of the P movement between adjacent plots was included in the calculation. These low recovery rates indicate that a significant proportion of the applied P is lost either in surface runoff or leached below the plough layer. However, the Chang and Jackson analysis of the subsoil samples did not reveal any significant changes in any of the P fractions. The lack of P accumulation might be attributed to the use of an invalid sampling strategy, to the loss of P to deeper layers through preferential flow and to the possibility that some P has been lost through surface runoff. The highest rate of P fertilization (P2K) resulted in a decrease of the P buffer capacity of the soil (BC) and increased the EPC0 (i.e. P concentration in solution at which no net P sorption or desorption occurs) as derived from the sorption isotherms. The resulting low BC and high EPC0 (1.30 mg L-1) are in agreement with measurements made in other heavily fertilized soils in Finland. The lower fertilization rate (P1 treatment) did not have an observable effect on the BC or the EPC0 (0.33 mg L-1). The depletion of soil P in the P0 treatment caused a decrease on the EPC0 (0.12 mg L-1) but only a slight increase on the BC. The degree of P saturation (DPS) of the plough layer was calculated as the ratio between the sum of P extracted with NH4Cl, NH4F and NaOH and the P sorption capacity of the soil (PSC; i.e. sum of extractable iron and aluminium multiplied by a 0.5 index (α)). Despite the large P applications during 37 years, the DPS reached only 20% for the topsoil of the P2K treatment, which can be attributed to the high PSC of the soil. The 0.5 α value is an operationally defined parameter, widely used in international publications, which is supposed to represent the proportion of short-range ordered Fe and Al oxides that contribute to the PSC of the soil. Due to the young age of Finnish soils and their usually high organic matter content, a low crystallinity of the oxides is expected, which could result in a higher α value. To prove this hypothesis, a long term incubation was done as a replicate of other published experiments (250 days, soil solution 1:50, P concentration 155 mg P L-1). Since the incubation includes only 4 samples from a single field, they cannot be extrapolated to other Finnish soils. However, the results (α =0.3) might be indicative of a low reactivity of the Al and Fe oxides in our soil. Iron and aluminium oxides are responsible for the accumulation in the soil of large amount of the applied P. However, the P recovery rates in field experiments are usually far for complete. To quantitatively assess the losses of P from the plough layer, a more congruent sampling strategy in the soil profile is needed, together with the analysis of P surface runoff and leaching wate