Organic P in temperate forest mineral soils as affected by humus form and mineralogical characteristics and its relationship to the foliar P content of European beech

Highlights

• Organic P and microbial biomass P were not affected by the humus form.

• Organic C to organic P ratios were lower in mull than in moder/mor soils.

• Organic P was positively related to clay and oxalate-extractable Fe and Al.

• Foliar P of beech could be predicted by the combined effect of organic P and clay.

Abstract

Phosphorus (P) deficiency of beech (Fagus sylvatica L.) is widespread in Europe. However, it is difficult to determine at which sites trees are prone to P deficiency, since foliar analysis is the only tool widely accepted for assessing the P nutritional status of beech. Our central objective was to identify the main influencing factors on organic P in forest soils and to test its usefulness as a predictor of the foliar P content of beech. Our hypotheses were: (1) stocks of total organic P (TOP) and microbial biomass P (MBP) in the mineral soil are higher at mull than at moder/mor sites, (2) the TOP content and the soil organic carbon (SOC):TOP ratio are affected by soil mineralogical characteristics, such as oxalate-extractable iron and aluminum (Fe_ox, Al_ox) and clay, (3) multiple regression models that include the TOP stock, soil mineralogical parameters and the humus form as predictors are useful to explain the variability of the foliar P content of beech. We sampled the mineral soils of 9 mull and 11 moder/mor sites at 0–50 cm depth and took foliar samples of beech trees in three subsequent years. In contrast to our hypothesis (1), we did not find significantly higher TOP or MBP stocks, but determined lower SOC:TOP ratios at mull compared to moder/mor sites. In line with our hypothesis (2), the TOP content was positively related to Fe_ox, Al_ox and clay contents, while the SOC:TOP ratio was negatively related to Fe_ox, Al_ox and clay contents, suggesting strong effects of clay, Al and Fe on the accumulation of TOP. In partial agreement with our hypothesis (3), the TOP stock was included as a positive and the clay content as a negative predictor in a multiple regression model of the foliar P content, suggesting decreasing bioavailability of TOP with increasing clay content. Across all sites, the model consisting of these two parameters better explained the foliar P content (R² = 0.60) than the stocks of Bray-II- (R² = 0.37) or Olsen-extractable P (R² = 0.53) as single predictors. Overall, our results suggest that the accumulation of TOP in forest mineral soils and its potential availability to beech are strongly affected by soil mineralogical parameters, which may be included in statistical models to improve the predictability of P availability to trees.

Introduction

Phosphorus (P) deficiency of beech (Fagus sylvatica) is widespread in Europe (Jonard et al., 2014; Talkner et al., 2015). According to the second German national forest soil inventory (conducted 2006–2008), foliar P contents of beech were below the normal range at 60% of the plots investigated (Riek et al., 2016). However, it is difficult to determine whether trees are prone to P deficiency at a given site, since foliar analysis is currently the only widely accepted tool available for assessing the P nutritional status of beech (e. g. Braun et al., 2010; Jonard et al., 2014; Merino et al., 2008). To individually assess the sustainability of forest management practices, it would be favorable to be able to predict P availability to trees from simple soil P tests or other soil parameters.

Various extractants for P have been tested to predict the P nutritional status or the growth response to P fertilization of several tree species under different climatic conditions. Generally best results have been obtained when the acid ammonium fluoride solutions of Bray and Kurtz (1945) or the alkaline NaHCO₃ solution of Olsen et al. (1954) were used (Ballard and Pritchett, 1975; Ballard, 1974; Mason et al., 2010; Radwan et al., 1985; Wells et al., 1986). However, these methods did not always prove successful (Kadeba and Boyle, 1978; Mendham et al., 2002) and have only rarely been tested in temperate forest soils. A major issue of these and most other classical soil tests is that they neglect organic P.

In forest soils low in total P (TP), the supply of P to trees is dominated by biological processes related to the turnover of organic P rather than by physicochemical exchange and dilution processes (Achat et al., 2009, Achat et al., 2012, Achat et al., 2013; Bünemann et al., 2016; George et al., 2017; Lang et al., 2017). To characterize the bioavailability of organic P, the sequential fractionation method of Hedley et al. (1982) has often been applied to forest soils (Johnson et al., 2003). However, organic P fractions separated by this method are defined operationally and the fractions considered non-labile may still be important for tree P nutrition (Fox et al., 2011; Niederberger et al., 2017; Richter et al., 2006). In contrast, methods for determining total organic P (TOP) are less laborious and do not separate different fractions (Turner et al., 2005). Despite this, TOP has been found to be closely related to the P uptake of plants for some soils (Harrison, 1987). For instance, Adepetu and Corey (1976) found that the TOP content of 21 Nigerian rainforest and savanna soils was a better predictor of the P uptake of maize than the contents of total P (TP) or Bray-I-extractable P. However, to our knowledge, it has never been tested whether TOP is a useful predictor of the P nutritional status of trees in temperate forest ecosystems.

The usefulness of TOP as a predictor of the P nutritional status of trees may be limited by the fact that TOP contents in forest soils are strongly affected by mineralogical characteristics such as the contents of clay or reactive aluminum (Al) and iron (Fe) (Grand and Lavkulich, 2015; Talkner et al., 2009; Vincent et al., 2012; Werner et al., 2017). This is probably due to a reduction of the microbial degradability of organic P compounds when interacting mineral components are present (Celi and Barberis, 2005). When investigating the P nutritional status of trees in response to the TOP stock, such a reduced bioavailability of TOP may be corrected for by including soil mineralogical parameters as independent variables in multiple regression models. Using a similar approach, Achat et al. (2013) were able to predict P mineralization from the ratio of organic matter to the sum of oxalate-extractable Al (Al_ox) and Fe (Fe_ox) ions in French Arenosols and Podzols.

TOP stocks in forest mineral soil layers may also be affected by factors related to the presence of different humus forms. Due to a higher degree of perturbation in mull compared to moder/mor soils, a higher proportion of litter is incorporated into the mineral soil, which may lead to an increased accumulation of soil organic carbon (SOC) (Andreetta et al., 2011; De Nicola et al., 2014) and possibly also of TOP (Harrison, 1979) in this compartment. Additionally, more favorable conditions for soil microorganisms might result in higher microbial biomass P (MBP) stocks in mineral layers of mull compared to moder/mor soils (Zederer et al., 2017). The turnover of organic P in the mineral soil might be more important for tree P nutrition at mull sites than at moder/mor sites, since trees can take up a considerable proportion of their P requirements from the forest floor when moder/mor type humus is present (Brandtberg et al., 2004; Jonard et al., 2009; Paré and Bernier, 1989a, Paré and Bernier, 1989b). Thus, when predicting the P nutritional status of trees from the TOP stock in mineral soil, the humus form might be another important independent variable.

Based on these considerations, our central objectives were to identify the main influencing factors on the organic P stock in mineral forest soils and to test its usefulness as a predictor of the foliar P content of beech. Our hypotheses were: (1) as a result of stronger perturbation, SOC, TOP and MBP in the mineral soil are higher at mull than at moder/mor sites, (2) the TOP content and the SOC:TOP ratio are affected by soil mineralogical characteristics such as Fe_ox, Al_ox and clay, (3) multiple regression models that include the TOP stock, soil mineralogical parameters and the humus form as predictors are suitable to explain the variability of the foliar P content of beech.