< Back to previous page

Project

Enhanced phosphate availability in weathered soils by rotations with legumes: a mechanistic analysis of P-fluxes in soils as measured by the DGT (Diffusive Gradient in Thin films) technique.

Phosphorus availability measurements using the diffusive gradients in thin films (DGT) technique in highly weathered soils</>

Phosphorus (P), next to nitrogen, is the most important limiting factor for crop production in many tropical and subtropical soils. Besides low total and/or available P contents, these soils are often highly P-fixing due totheir high aluminium and iron oxide concentration (Norman et al. 1995; Sanchez et al. 1997). Essential for the evaluation of different plant-soil systems is the correct determination of plant-available P and profound knowledge about P dynamics. The starting point for correcting the soilfertility status is soil testing. Over the past century, many soil P tests have been established and put forward as being the ultimate index ofP availability in soils. Unfortunately none of these techniques seems adequate to correctly assess the phosphorus status in strongly weathered acid soils as found in tropical areas. These established soil P tests are based on chemical extraction with a strong base, acid or complexing agent. Typically soil tests only give reliable results for the soils they were calibrated for (Degryse et al. 2009b). The research presented aims at investigating whether the diffusive gradients in thin films (DGT) technique allows a better prediction of available P compared to establishedsoil P tests.

The DGT technique is a passive sampling techniquethat has been successfully applied to aquatic systems for measurement of trace metal concentration (Zhang and Davison 1995; Zhang et al. 1998b), later for phosphorus measurements (Zhang et al. 1998a) and more recently for predicting crop response to applied P in a wide range of soils (Mason et al. 2010; McBeath et al. 2007; Menzies et al. 2005; Tandy et al.2011). In contrast to previous studies, this study focuses on strongly weathered, acid, tropical soils. Similar to the anion exchange membrane methods (AEM), the DGT technique attempts to mimic the physico-chemical uptake of P by plant roots by providing a sink for free phosphate (Masonet al. 2010). The continuous removal of P from solution lowers the concentration of P in the soil solution (intensity factor, I), which in turnpromotes resupply from the soil solid phase, depending on the labile P pool or P quantity of a soil (Q) and the resupply capacity, P buffering capacity (PBC). Provided that diffusion of P towards plant roots is the rate limiting step for P uptake under P deficient conditions, the DGT measured concentrations, which are a measure for such a diffusive flux andP intensity, have potential to be a better predictor for P availabilityto plant roots than established soil P tests, which are rather measuresfor P quantity in a soil.

Firstly, it was hypothesized that soilP tests that extract P from the plant accessible pool, will predict availability and uptake more robustly than empirical tests, i.e. based on chemical extraction. This was tested by comparison of the isotope ratios (33P/31P, specific activity, SA) of P between plant shoot and the soil extract. When the SA of the plant shoot is not significantly different from the SA in the soil extract, the same P pool is most likely advocated.In contrast, when the SA in the soil test is smaller than that in the maize shoot, P was extracted from a P pool not accessible to maize. To this test we submitted the DGT technique, in comparison with conventional soil P tests viz. Olsen, Colwell, Bray-1, Mehlich-3, ammonium oxalate, AEM and 0.01 M CaCl2 solution. A pot trial was conducted with maize as a test crop, grown in two P deficient soils from western Kenya with contrasting P sorption characteristics, amended with a low and a high P rate and labelled with 33P. A close correspondence was found between the SA ofthe extracts of the different soil tests (except CaCl2 and ammonium oxalate extracts) and that of the plant, at the low P rate in the soil withlow P sorption capacity. For the high P rate on this soil, differences in SA between maize shoot and soil test were small for all established soil tests, but significant for the Colwell, Bray-1, Mehlich-3 and AEM test. The SA in the soil extracts was significantly smaller than that in the maize shoot for the strongly P-sorbing soil at both P rates for all conventional tests, including AEM. This indicates that these tests extracted P from a pool that is not accessible to the plant. For the DGT test,however, there was no difference in SA between the maize shoot and the soil test, for any of the treatments. From this experiment, we could conclude that most conventional soil tests extract a fraction of P which isnot available to maize. In contrast, the DGT technique samples only P from the plant-accessible pool.

Secondly, a pot trial was conducted to test DGT relative to established soil P tests in predicting the growth response to P addition across nine different tropical soils marked by P deficiency, for two plant species with contrasting growth rate affecting P demand per dry matter unit, i.e. maize and upland or rainfed rice. Data from previous pot and field trials support our hypothesis that the predictive power of the DGT is superior compared to established soil P tests, because DGT provides a measure for the P supply to plant roots under P deficient conditions where the rate limiting step for P uptake is diffusional supply of P to the roots. We found that the DGT method andCaCl2 extractions explained relative yield (% of maximum yield) of maize among soils better (R² = 0.82 and 0.75 respectively) than P determinedby Olsen, Colwell, Bray-1, Mehlich-3, ammonium oxalate and resin extractions (R² < 0.53). In strong contrast, relative yields of rice were bestpredicted by Mehlich-3, Bray-1, Olsen and resin P (R² ~ 0.7) compared to DGT (R² = 0.53) and CaCl2 (R² = 0.08). Moreover, modelling results by Degryse et al. (2009b) suggest that the critical DGT concentrations, i.e. the CDGT to obtain 80 % of the maximum yield, are plant specific rather than soil specific. Our results confirm this suggestion: the critical DGT P concentrations on this set of soils were 73 µg P L-1 for maize andonly 7 µg P L-1 for rice. The critical DGT value measured for maize is in correspondence with literature. In short, for tropical, P deficient soils, intensity-based indices of soil P availability such as DGT and CaCl2, are superior to quantity-based indices (i.e. the established soil P tests based on extraction) for maize with high P demand. However, the reverse is true for rice suggesting that diffusion of P in the soil as measured by DGT is not the main factor explaining P uptake for rice. We suggest, but do not prove, that this upland rice variety disposes of strategies to overcome P deficiency (e.g. exudation of organic acids or phosphatases, changing root architecture).

Finally, our goal was to evaluate the potential of the DGT technique to predict bioavailable P in different soils amended with various qualities of OM. This was investigated since it is known that the P fertilizer value of organic materials (OM) such as plant residues or manure, can be larger than that of mineral fertilizers in weathered soils. In our final experiment, phosphorus was applied at various rates (deficient to adequate) as triple superphosphate (TSP) only or at one intermediate P dose, in a substitution trial with4 different OMs: farmyard manure (FYM) or residues of Tithonia diversifolia, each at low or high P content. Results demonstrated that the combined application of TSP and OM increased, decreased or did not affect drymatter yields compared to single TSP application at corresponding totalP dose, differences depending on soil and OM type. Relative yield to soil P test values correlated most strongly for DGT measured P concentrations, R² is 0.74, compared to Olsen (R² = 0.60) and AEM (R² = 0.62), whenconsidering all treatments with a P-related effect. This selection was needed because also non P-related factors as a consequence of OM application, can have an effect on growth. Results of this study suggest that DGT performs better in detecting TSP × OM interactions in different soilsthan the established soil P tests (Olsen and AEM); although not all trends in DM yields can be fully understood using the DGT technique.

In conclusion, this work demonstrated the strong prediction power of DGT in P deficient soils for maize yields, compared to established soil Ptests. This strong prediction power is not surprising, given that DGT samples P from the plant accessible pool of P in the soils, while with the conventional soil P tests P was also extracted from a pool not accessible to plant roots, which must lead to an overestimation of plant available P. For rice the prediction power of quantity measures seems larger, which can most likely be related to their strategies to overcome P deficiency. Our results confirm that the critical DGT concentrations are not soil dependent, but rather are plant specific. Finally, the critical DGTconcentrations for maize were constant over the two response experiments, even when different growing periods, growth conditions, P amendments or soils are used. Our work demonstrates that the DGT technique is a robust method to assess the P status of a soil and to correct it if required.

Date:1 Oct 2008  →  21 Dec 2012
Keywords:Weathered soils, Phosphate
Disciplines:Agricultural animal production, Agricultural plant production, Agriculture, land and farm management, Other agriculture, forestry, fisheries and allied sciences
Project type:PhD project