Soil Phosphorus Fractions as Influenced by Different Cropping Systems: Direct and Indirect Effects of Soil properties on Different P Pools of Nitisols of Wolayta, Ethiopia

Data from 12 surface soils (0 – 15 cm depth)of three cropping systems (enset, maize and grazing land) and path analysis was used to evaluate effects of soil properties: pH, texture (Clay, silt and sand) , organic carbon (OC) cation exchange capacity (CEC),citrate-dithionite-bicarbonate (CDB) extractable Fe and Al (Fed and Ald) on total phosphorous (Pt), organic phosphorous (Po), Olsen P (Available P) and Chang and Jackson (1957) inorganic phosphorous (Pi) fractions. Correlation analysis was performed to study the relationships between soil properties and different soil P pools while path analysis model was used to evaluate direct and indirect effect of these soil properties on the P pools. Only soil properties that significantly contribute to the fit of the model were used. High significant values of coefficient of determination (R2) and low values of uncorrelated residual (U) values indicate the path analysis model explains most of the variations in soil Pt, Po, Olsen-P, Saloid-P, Ca-P, Al-P, and Fe-P pools. Soil pH had significantly high and positive direct effect (D = 0.618*) on Pt, (D = 1.044***) on saloid P, and (D = 1.109***) on Fe-P with modest and negative indirect effect (D= -0.478 and -0.405) on saloid P and Fe-P, respectively, through OC. The direct effect of clay on Ca-P, Al-P and Fe-P (readily available P forms) was significant and negative with a relatively higher indirect effect on Fe-P through pH suggesting that clay is dominant soil property that influences readily available P pools in Nitisols of the study area. Fed had significant and negative direct effect (D = -0.430*) on Olsen available P with low negative indirect effect ( D = -0.154) through pH results in significant and negative correlation (r = -0.657*). The significant and negative direct effect of Fed on Olsen P indicates that crystalline iron is the sink for available P. Relative influence of the soil properties on the soil P pools was in the order: pH > clay > Fed > OC.  These results show that most of P pools of Nitisols of Wolayita are best predicted from pH, clay (texture), Fed and OC. On the other hand, our data also show that the inclusion of other soil variables is needed to fully predict Ca-P and stable P pools