5.1 Overview of the Phosphorus Cycling in the Soil–Plant Continuum

Phosphorus (P) is the 11th most abundant element in the Earth’s crust, which is equivalent to around 0.12 %. The amount of P in the bedrocks varies (Van Kauwenbergh 2010), with the highest amounts usually found in sedimentary rocks. As bedrock is the primary source of mineral P in soils (Ruttenberg et al. 2003), the bedrock type and composition influence the total amount of P in different soil types. Total P concentrations in natural soils are usually lower than in the original bedrock material, with higher concentrations in younger than in older soils, which also differ in their relative abundance of P forms (Walker and Syers 1976). Generally, calcium-bound P dominates in younger soils, whereas occluded inorganic P and organic P forms become more abundant with progressing soil development. P concentrations and the relative proportion of various P forms vary in soils with age and have been studied extensively since the pioneering work of Walker and Syers (1976). The most recent study by Rechberger et al. (2021) examined the sorption and desorption of phosphate in volcanic soils on the Galápagos Islands. Their study sites included a climatic gradient, with annual rainfall ranging between 100 mm (dry sites) and 1600 mm (humid sites). They concluded that the main drivers for P sorption at their sites were iron (Fe) and aluminium (Al), either in amorphous phases or bound to organic matter.

The following paragraphs provide an overview of P cycling in the soil–plant continuum, for more detailed information see, for example, Arai and Sparks (2007), Raghothama and Karthikeyan (2005), White et al. (2008), Bünemann et al. (2011), Shen et al. (2011), Weihrauch and Opp (2018), and Schipanski and Bennett (2021).

5.1.1 Phosphorus Cycling in the Soil

Through weathering processes, P is released from the bedrock into the soil where it is transformed into the different organic and inorganic P forms and taken up by the soil biota. Weathering processes are divided into physical and (bio-) chemical weathering processes, and factors controlling weathering (such as climate, soil type, mineralogy, and other intrinsic soil properties) are highly inter-related (Scheffer and Schachtschabel 1998). Chemical weathering, for example through hydrolysis of minerals and oxidation of ions, of bedrock dominates in humid tropical regions, whereas physical weathering processes like thermal stress and frost weathering are more important in arid areas. The biochemical weathering of bedrock is favoured by processes such as exudation of organic acids by bacteria and plants and acidification due to the excretion of protons.

The importance of bedrock weathering for the P cycle in the soil–plant continuum is illustrated in the following paragraph. For more information see, for example, Formoso (2006), Derry (2009), Goll et al. (2014), and Ribeiro et al. (2020) and references within. Soil water content is used as an example in the following paragraph to illustrate the interconnectivity of the factors controlling weathering and how it influences the cycling of other P forms in the soil–plant continuum (Arai and Sparks 2007) (Fig. 5.1).

Fig. 5.1
A schematic depicts the interconnectivity of the factors controlling weather and the complexity of soils. It begins with cloudy weather condition that causes soil mineralogy, soil water content, soil biotic processes, soil aggregation, and diffusion rate.

Schematic diagram showing the complexity of soils

Full size image

We chose the soil water content as an example as it impacts the diffusion rate of ions in the soil which in return impacts the adsorption rate of ions like phosphate (Weihrauch and Opp 2018). It also enhances the chemical weathering of bedrock material and impacts biochemical weathering, as it directly affects soil fauna and flora. Thus, soil water content is an important factor for P cycling in the soil–plant continuum. It is impacted by factors such as precipitation and temperature (climate), soil mineralogy, and soil biotic processes. Precipitation and temperature also affect soil biotic processes (Krashevska et al. 2012; Margalef et al. 2017; Nottingham et al. 2020), the P cycle (Hou et al. 2018), and influence changes in soil mineralogy (Boero et al. 1992). Through their burrowing activities, earthworms ingest soil and via this and their excretions, they also influence soil aggregation (Blouin et al. 2013), which in return influences soil microorganisms and P sorption (Linquist et al. 1997; Six et al. 2004). Earthworms, soil fauna in general, and roots create biopores through their activities, which impact soil porosity and nutrient distribution in soils (Le Bayon et al. 2021). In return, soil fauna and flora influence bedrock weathering, as well as the P cycle. Through the exudation of enzymes by soil fauna and flora, organic P is hydrolyzed and available for plant uptake.

Similarly to weathering, chemical and physical processes (including precipitation, dissolution, adsorption, desorption) as well as biological processes (including immobilization, mineralization, enzymatic hydrolysis, and the activity of soil macro-, micro-, and mesofauna and -flora) govern the P cycle in the soil–plant continuum (Fig. 5.2) (Arai and Sparks 2007; Vos et al. 2014).

Fig. 5.2
A schematic diagram depicts the P cycling. The phosphate in soil solution includes mineralization, immobilization, uptake to soil fauna, weathering, and uptake to plat residues.

Schematic of phosphorus (P) cycling in the soil–plant continuum

Full size image

Phosphorus occurs as inorganic and organic P forms (e.g. phosphate, phosphonates, and phytate) in soils, with inorganic and organic phosphates being the most abundant (McKelvie 2005; Arai and Sparks 2007; Tiessen 2008). The bioavailability of soil P varies greatly, and some P forms such as phytate are more recalcitrant than others. Only a small amount (around 0.1 mg P L−1; Tiessen 2008) of inorganic P occurs as free phosphate, i.e., the readily available P, in the soil solution and can directly be taken up by plants. Phosphate in soils is most available at a neutral pH range but this range is recently under discussion (Barrow 2016; Simonsson et al. 2018). Most inorganic P occurs as phosphate bound to aluminium (Al) or iron (Fe) oxyhydroxides, to soil particles like clay minerals, and to calcium (Ca) or magnesium (Mg) (Gérard 2016). Soil microorganisms like archaea and bacteria also partly contain inorganic P (around 30 %; Bünemann et al. (2011)), but the majority of P in microorganisms occurs as organic P, such as phospholipids and nucleic acids. The relative proportion of organic P forms in microorganisms is influenced, similarly to plants (Veneklaas et al. 2012), by their P status. In P-limited aquatic bacteria, 60% of microbial P consisted, for example, of phospholipids and nucleic acids (Vadstein 1998). Microorganisms also often store P in the form of polyphosphates in their cells (Rao et al. 2009; Akbari et al. 2021). Depending on the soil, organic P typically makes up 30 to 65 % of total soil P (Harrison 1987; Richardson et al. 2005).

5.1.2 The Fate of Fertilizer P in Soils

As concentrations of bioavailable P in agricultural soils are often insufficient to maintain crop production, farmers apply P fertilizer (Schnug and De Kok 2016; Ros et al. 2020). Phosphorus fertilizers are available as mineral and organic fertilizers. Mineral P fertilizer mainly consists of inorganic P, whereas organic P fertilizers contain inorganic P but also varying amounts of organic P compounds and other organic compounds, such as humic acids. Due to their different compositions, mineral and organic P fertilizers behave differently in soils (Audette et al. 2016) and affect P cycling in the soil–plant continuum in different ways (Keller et al. 2012; Ma et al. 2020). Adding organic P fertilizers to the soil leads to an increased input of organic matter compared to mineral P fertilizers. This can result in enhanced microbial activity in organically fertilized soils (Oehl et al. 2001). Mineral P fertilization on the other hand can lead to reduced microbial and enzymatic activity in soils (Zhang et al. 2015; Chen et al. 2019).

5.1.3 Mineral P Fertilizers

Rock phosphate, derived from P-containing ores, is the primary raw material for mineral P fertilizers, but it can also be applied directly to agricultural fields (Zapata and Roy 2004). Mineral P fertilizers are applied in agricultural fields as triple superphosphate (TSP), (mono-/di-) ammonium phosphate, combined with other nutrients such as nitrogen (N) and potassium (K), and to a lesser degree as ammonium polyphosphate, which is, unlike the first three, applied as a liquid. Only a small amount (up to 20%; Plaxton and Tran (2011)) of inorganic P derived from mineral fertilizers is taken up by plants during a growing season (Blake et al. 2000). The majority is either adsorbed by soil particles and slowly enters the more recalcitrant P pools or is lost due to leaching into the groundwater or soil erosion. In developed countries with access to fertilizers, this process translated into the build-up over a few decades of the so-called legacy P in soils (Kamprath 1967; Menezes-Blackburn et al. 2018). Ott and Rechberger (2012) estimated that in the 15 European Union Countries as of 1 May 2004 (EU-15), the net accumulation of P in agricultural soils is 2.9 kg P yr−1 per capita. This is approximately 62% of the net annual per capita consumption of P in the EU-15.

Organic P amendments cover a wide range of substances. Farmyard manure is probably the most common organic P amendment; other examples are biogas digestates and animal slurry. The origin of the organic P fertilizer/amendment impacts the fate of P derived from it, as shown for three different types of animal manure (Azeez and van Averbeke 2011) and a range of organic fertilizers including digestates (Vanden Nest et al. 2015). Poultry manure, for example, reduced P fixation in soils more than goat and cattle manure, perhaps caused by a competition about the sorption sites between P and humic acid (Azeez and van Averbeke 2011). Similarly, Vanden Nest et al. (2015) observed that the different compositions of organic matter in organic fertilizers lead to different effects on soil P availability. The various organic P fertilizers also contain different organic P compounds or different amounts of organic P compounds. As organic P compounds need to be hydrolyzed before they can be used by plants (see Sect. 5.3), this also impacts the availability of P derived from organic P fertilizers. Phytate is, for example, more present in poultry manure and pig slurry, as these monogastric animals lack the enzyme phytase to hydrolyze phytate (Scholz and Wellmer 2015).

5.1.4 Phosphorus and Plants

The amount of total P in soils would often be sufficient to sustain plant growth, however, only a small portion of the total soil P can readily be taken up by plants or soil microorganisms. Plants take up P as a phosphate ion, mainly as H2PO4− or HPO42−, from the soil solution, creating a P concentration gradient from the rhizosphere to the bulk soil, which leads to the desorption of phosphate from soil particles and a flux of phosphate towards the roots (Hinsinger 2001; Schnepf et al. 2008; Kreuzeder et al. 2018). Since concentrations of available P in soils are usually low compared to plant P demands, plants have developed strategies to increase their P acquisition and internal P use efficiency (Table 5.1) (Hinsinger 2001; Shen et al. 2011; Veneklaas et al. 2012; Raven et al. 2018).

Table 5.1 Plant strategies to increase phosphorus (P) acquisition and internal P use efficiency
Full size table

Acquisition strategies can roughly be divided into four categories: exudation, symbiosis, expression of high-affinity P transporters, and root architecture. Plants are known to excrete organic acids, enzymes, and protons to increase the availability of soil P. Organic acids chelate ions like Fe and thus prevent the binding of P to Fe. Pigeon pea (Cajanus cajan) exudes piscidic acid (p-hydroxybenzyl tartaric acid), which is known to release P bound to iron (Fe) (Ae et al. 1990; Raghothama 1999; Adu-Gyamfi 2002; Ishikawa et al. 2002; Krishnappa and Hussain 2014). Enzymes such as acid phosphatase catalyze the hydrolysis of organic P into inorganic P so that plants can take it up. As mentioned in Sect. 5.1, the soil pH affects the availability of the soil P, and therefore some plants excrete protons, which leads to an acidification of the rhizosphere. Other plants express high-affinity P transporters in the roots, which leads to an increased P flux into the roots (Rausch and Bucher 2002). Changes in root architecture can also increase P acquisition. Members of the Proteaceae species form cluster roots that excrete organic acids. The formation of proteoid (or cluster) roots is, for example, a common strategy of the Proteaceae family in Australia growing on low P soils (Lambers et al. 2015a). Among the cultivated species, white lupin (Lupinus albus) is also known to form cluster roots (Lambers et al. 2015b). Phosphorus-limited plants often also invest more resources in the roots, which leads to a greater root-to-shoot ratio compared to non-limited plants. An increased number of fine roots allows plants to explore a greater soil volume. In most soils, bioavailable P concentrations are highest in the topmost soil layers. A strategy to better access this P is a change in the root architecture, with more roots growing laterally (Lynch and Brown 2001). This phenomenon is known as topsoil foraging and is increased in soybeans by a shallower root growth angle (Lynch 2011). Similarly, the symbiosis with mycorrhizal fungi greatly increases the soil volume that plants can explore and consequently have access to more bioavailable P (Schnepf et al. 2008; Smith et al. 2011). Mycorrhizal fungi, like plants and microorganisms, secrete phosphoenzymes which could increase the availability of organic P for plant nutrition (Ezawa and Saito 2018). In fact, around 90% of land plants are associated with mycorrhizal fungi (Smith and Read 2010).

Phosphorus is essential for plants as it is involved in several metabolic processes and the inorganic P concentration inside the cytoplasm is tightly regulated (Bieleski 1972; Raghothama and Karthikeyan 2005). Therefore, plants, besides their P acquisition strategies, have developed mechanisms to increase their internal P use efficiency (Plaxton and Tran 2011). More detailed information about this topic can be found in White and Hammond (2008), Rose and Wissuwa (2012), Lopez-Arredondo et al. (2014), and references therein. Inside plants, P is considered a mobile nutrient, and once it is taken up by plants, it is translocated to the different plant organs. The uptake of P and translocation to different plant organs and cell compartments involve cell membrane transporters and the xylem and phloem (Jaiwal et al. 2008; Miller et al. 2009; White 2012). Depending on the P nutrition status of a plant, most of the plant P occurs as inorganic P (Veneklaas et al. 2012) mainly stored inside the vacuole when plants are growing at a high P status (up to 85–95 %; Raghothama and Karthikeyan (2005)). The remobilization of this stored vacuolar P is one way in which plants can increase their internal P use efficiency (PUE) (Rose and Wissuwa 2012). Re-translocation of P from older to younger or reproductive organs like seeds is another strategy to increase the PUE (Akhtar et al. 2008; Richardson et al. 2011). Compared to the tightly regulated cytosolic inorganic P, the concentrations of other plant P compounds like nucleic acids, phosphorylated metabolites, and phospholipids can change more and the turnover rates of some of these P compounds are less than 1 min (Bieleski and Laties 1963; Veneklaas et al. 2012). After the addition of 32P to potato plants, it took less than 1 min until half of the ATP in older potato tissue had incorporated the added label (Bieleski and Laties 1963). The replacement of phospholipids by other lipids like sulfo- and galactolipids also increases the PUE of plants (Lambers et al. 2012). Similar to increasing the P acquisition, the expression of intracellular acid phosphatases increases the PUE of plants (O’Gallagher et al. 2021).

Most modern crop varieties are less efficient in acquiring P compared to older varieties or wild types since breeding efforts focused on other traits like increased starch content or yield in general. Nowadays, researchers have investigated the P uptake and utilization efficiency of different plant varieties, and through breeding or genetically modifying crops to create new crop varieties, which have a good balance between yield and P use efficiency (Cong et al. 2020). Examples include potatoes (Balemi and Schenk 2009; Pantigoso et al. 2020), rice (Wissuwa et al. 1998; Irfan et al. 2020), and wheat (Fageria and Baligar 1999; Korkmaz et al. 2009). With these new varieties, it might be possible in future to reduce the dependence on P fertilizers, as well as reduce the amount of legacy P built up in the soil. Other research focuses on soil management strategies such as intercropping (Brooker et al. 2015; Garland et al. 2017) and microorganisms or soil meso- and macrofauna to increase P availability in the soil (Richardson et al. 2009; Blouin et al. 2013; Trap et al. 2016). Intercropping cereals with legumes can have a positive effect on the P availability of cereals (Hinsinger et al. 2011). When intercropping maize (Zea mays) with fava beans (Vicia faba), maize as well as fava bean yield increased by 43% and 26%, respectively (Li et al. 2007).

5.1.5 The Next Steps—Where Are the Biggest Gaps in Our Knowledge?

As the previous section illustrates, P cycling in the soil–plant continuum is complex and we still do not fully understand all its aspects. At the same time, dealing with P-related issues (such as declining rock phosphate reserves and pollution of aquatic systems) becomes more urgent. Several research papers recently addressed research gaps and questions regarding the P cycle in the soil–plant continuum (Yang and Finnegan 2010; Richardson et al. 2011; Kruse et al. 2015; Reed et al. 2015; George et al. 2017). Generally, the gaps can be divided as follows:

  1. i.

    Soil intrinsic factors, management, and fauna. Organic soil P can be an important P source for plants; however, many issues around organic P remain, for instance how to link enzymatic P cycling with operationally defined (organic) P pools or the importance of microbial mechanisms related to organic matter (George et al. 2017). The role of soil pH regarding P availability has recently been under discussion (Barrow 2016; Simonsson et al. 2018). It is also still unclear how microorganisms deal with low P conditions at acidic soil pH (Lidbury et al. 2017). Like soil pH, soil management can also affect the availability of soil P. Soil management strategies include intercropping different plant species with each other and adding specific microorganisms to the soil to enhance P availability. Despite recent advances in this field, there are still many unknowns about how soil management strategies can increase the availability of legacy P and other recalcitrant P (Menezes-Blackburn et al. 2018). Soil management also affects earthworms and other soil fauna, but the role of earthworms for the P nutrition of crops, for example, is a relatively recent field of research (Puissant et al. 2021; Trap et al. 2021).

  2. ii.

    Plant physiology and the genetics behind. Whereas many studies, mainly under experimental conditions in the greenhouse, have examined how P starvation affects plant physiology and what genes are involved, it is still unclear what sets off the phosphate starvation response of plants (Yang and Finnegan 2010). It is also debated how relevant are data obtained in greenhouse studies for field-grown plants (Richardson et al. 2011). Greenhouse studies usually only consider a few variables like the soil P status or soil water content. In contrast, plants grown under field conditions might not only be exposed to low P concentrations but also must adapt to changes in precipitation, temperature, competition about P with microorganisms, or a heterogeneous distribution of P in the soil.

  3. iii.

    Incorporation of P into environmental models. Improving the representation of the P cycle in models can improve the constraints of the predictions about soil C storage under climate change (Helfenstein et al. 2020). However, many model parameters regarding the P cycle are still missing (Reed et al. 2015). One aspect that is recently of high interest is linking the P to other nutrient cycles since modellers have realized the interconnectivity of nutrients (Wang et al. 2007; Reed et al. 2015; George et al. 2017; Bertrand et al. 2019). The C:N:P ratio is an important parameter for investigating the coupling of the C, N, and P cycle (Bertrand et al. 2019). Phosphorus is involved in key metabolic processes, such as providing energy for the transport of nutrients across membranes, and thus, P limitation can also affect the cycling of other nutrients. Cleveland et al. (2002) showed that in tropical forests, the utilization of labile soil organic carbon by microorganisms is P limited.

5.2 Current Approaches and Methods to Study P Cycling in the Soil–Plant Continuum

There are still many unknowns when it comes to P cycling in the soil–plant continuum. A detailed overview of current innovations regarding methods to investigate P cycling in the soil–plant continuum can be found in Bünemann et al. (2011), the IAEA TECDOC-1805 (IAEA 2016), Kruse et al. (2015), Neumann et al. (2009), and references therein. Methods to investigate P cycling in the soil–plant continuum can be divided into four groups: chemical extractions, tracers, molecular approaches, and spectroscopic methods (Bünemann et al. 2011). These methods vary in their areas of application and in the information obtained from them (Table 5.2). The chosen method therefore depends on the research questions and hypotheses. However, none of those methods should be used as a stand-alone method but should be combined with other approaches.

Table 5.2 Common methods to investigate P cycling in the soil–plant continuum
Full size table

Recent developments in the suitability of the different methods have driven attention to approaches commonly used. Recently, the usefulness of chemical fractionations to study P in soils has been debated (Klotzbücher et al. 2019; Barrow et al. 2021; Guppy 2021; Gu and Margenot 2021). In fractionation protocols, fractions are often assigned to certain soil P pools or certain P forms are assigned to certain fractions, like iron-bound P to the NaOH fraction. However, one should be aware that, for instance, a chemically determined P pool does not necessarily adequately represent an environmentally relevant P pool. The resin-extractable P fraction or pool is for instance considered as an approximation for available P, but other fractions also might contain available P. Some methods, like the δ(18O)PO4 technique, however, require knowledge about chemical fractionations (Tamburini et al. 2018), but the limitations of those should then be acknowledged when interpreting the results. It might be useful to combine chemical fractionations with other analytical methods, such as the determination of iron in the NaOH fraction, to validate the results obtained by chemical fractionations (Condron and Newman 2011).

The extraction and analysis of P fractions comprise a wide range of methods, from the routine extraction of available P in soils to the isolation of vacuolar P in plants. Most studies investigating P in the soil–plant continuum determine at least the easily extractable inorganic P (considered the available soil P) or the total P in soils. Common methods for the extraction of available P from soil samples include extraction with anion exchange resin membranes (Kouno et al. 1995), bicarbonate (Olsen 1954), water (Van der Paauw 1971), and a sodium fluoride-hydrochloric acid solution (Bray and Kurtz 1945). Not all methods are appropriate for all soil types (Oberson et al. 1997; Nawara et al. 2017; Blackwell et al. 2019), thus leading to erroneous results. Total P in soils and plant material is often determined via chemical extraction, such as acid digestion and sodium carbonate fusion (Sommers and Nelson 1972; Bowman 1988; Bender and Wood 2000; Maathuis 2013). Total P is also determined with less destructive methods like X-ray fluorescence (XRF) (see section below about spectroscopic approaches). While the extraction of available P from soils is relatively straightforward, extracting specific P compounds like phytate from soil and plant material is often more laborious (Frank 2013; Reusser et al. 2020; Turner et al. 2020). To quantitatively analyse phytate in soil and plant material, the samples were first extracted with an acid or alkaline solution, followed by isolating phytate from the extract. Methods for isolation include high-performance liquid chromatography (HPLC) and precipitating phytate as an insoluble salt, for example, by adding iron (Frank 2013; Turner 2020).

Using tracers to investigate nutrient cycling is a common research approach, and in the case of P, it includes the two radioisotopes 32P and 33P, the stable isotope composition of oxygen (O) associated with P, 13C-labelled organic P compounds, and using elements that are associated with P, such as uranium (U) in the case of fertilizers. The application areas of P radioisotopes are quite diverse (Frossard et al. 2011) and include determining the residence time of P in different soil P pools (Helfenstein et al. 2020) and tracing P inside plants once it has been taken up (Mimura et al. 1996). Recently, Whitfield et al. (2018) published a method to synthesize inositol hexakisphosphate (IHP) labelled with 32P, which could advance our understanding of IHP cycling in soils.

Using molecular approaches is becoming increasingly accessible. While only a few publications used omics analysis to investigate P cycling in the soil–plant continuum back in 2011 (Wasaki and Maruyama 2011), those methods are nowadays more common and they brought useful insights, for example, into the phosphate starvation response (PSR) of plants (Lopez-Arredondo et al. 2014; Lan et al. 2017). Proteomics revealed that the PSR of plants also affects the enzymes involved in the tricarboxylic acid cycle of plants (Lan et al. 2017).

Spectroscopic and spectrometric approaches cover a wide range of techniques (Kruse et al. 2015). Determining total P in soils via XRF instead of chemical extraction often yields more accurate results, as chemical extractions tend to underestimate the concentrations of total soil P (Chander et al. 2008; Wang et al. 2021b). The advantages and disadvantages of 31P-NMR are an ongoing debate. One disadvantage is that some organic P compounds might be hydrolyzed prior to or during NMR analysis. By using 18O-enriched medium this issue could potentially be solved (Wang et al. 2021a). Combining 31P-NMR with an 18O-label is also used to study enzymatic mechanisms (Cohn 1958; Cohn and Hu 1978). Spectroscopic approaches also include other approaches that are currently less common than those mentioned in Table 5.1 (Kruse et al. 2015). One example is nanoscale secondary ion mass spectrometry (NanoSIMS). Rodionov et al. (2020) used NanoSIMS in combination with other techniques to study P cycling in forest subsoils by visualizing the distribution of P in the rhizosphere.

5.3 The δ(18O)PO4 to Investigate P Cycling in the Soil–Plant Continuum—Current State

5.3.1 Background of the Method

In recent years, using the stable isotope composition of oxygen (O) associated with P [δ(18O)PO4] to study phosphorus (P) cycling in the soil–plant continuum has become more common. Detailed reviews about the principles of the method and its applications can be found in Adu-Gyamfi and Pfahler (2022), Bauke (2021), Tamburini et al. (2014b), and Jaisi and Blake (2014). In short, the principles of this method are (1) in the environment, P is mainly associated with O and (2) the P-O bond is stable under the absence of biotic processes and at the Earth’s surface conditions (Winter et al. 1940). Alterations of δ(18O)PO4 values in the environment occur due to processes that lead to cleaving of the P-O bond (mainly biotic processes) or sorting of the heavier and lighter isotopologues (biotic and abiotic processes) (Adu-Gyamfi and Pfahler 2022). Biotic processes include enzymatic activity, such as organic P mineralization through enzymes and P uptake by microorganisms. One of the most important enzymes influencing δ(18O)PO4 values is inorganic pyrophosphatase (PPase). This ubiquitous enzyme leads to the progressive exchange of all four O atoms in phosphate with O from surrounding water and causes a temperature-dependent isotopic equilibrium between O in phosphate and in water (Cohn 1958; von Sperber et al. 2017). Equilibrium δ(18O)PO4 values can be calculated (Chang and Blake 2015) and are often used as an indicator for microbial P cycling, since the PPase is an intracellular enzyme (Tamburini et al. 2012). Other enzymes, phosphomono- and -diesterases such as acid and alkaline phosphatases among others, lead to an exchange of one to two O atoms between the phosphate moiety and water (Liang and Blake 2006a; von Sperber et al. 2014; Wu et al. 2015). The reported fractionation factors are mostly negative; i.e., phosphate released from organic P is usually depleted in 18O compared to the original organic P compound. Acid phosphatase for example has a fractionation factor of approximately −10 ‰ (von Sperber et al. 2014). With a δ(18O)H2O value of the surrounding water of around 0 ‰ and organic P with a δ(18O)PO4 value of around 20 ‰, phosphate released from organic P via acid phosphatase would have a δ(18O)PO4 value of 10 ‰. Abiotic processes like the interaction of phosphate with iron oxides have a rather small fractionation factor compared to biotic processes (Jaisi et al. 2010; Melby et al. 2013a).

To determine the δ(18O)PO4 of a P fraction, the targeted fraction needs to be extracted from the environmental sample, e.g. soil, and afterwards isolated to precipitate the final analyte silver phosphate. Nowadays, protocols for the extraction, isolation, and purification of phosphate for determination of δ(18O)PO4 values exist for organic and inorganic P forms and fractions in soils, plants, sediments, fertilizers, and water (see Adu-Gyamfi and Pfahler, 2022 for an overview). Most protocols consist of a stepwise purification of an extract/sample (McLaughlin et al. 2004; Tamburini et al. 2010, 2018). The most common protocol for soils and plants consists of four steps: (1) precipitation of ammonium phosphomolybdate (APM), (2) dissolution of APM and precipitation of magnesium ammonium phosphate (MAP), (3) dissolution of MAP and addition of exchange resin to remove cations, (4) removal of resin and addition of silver ammine solution to precipitate silver phosphate. For a more detailed description of the method see also IAEA book, Tamburini et al. (2010), and Tamburini et al. (2018).

5.3.2 Calculations

To assure the quality of the obtained isotope data, the following calculations should be performed prior to further data analysis:

  • Estimation whether or not inorganic hydrolysis of organic or precipitated P occurred during the extraction of a P pool (McLaughlin et al. 2006; Pistocchi et al. 2017).

  • Recovery of a label like 18O-labelled phosphate or water added to soils, plants, or nutrient solutions (Gross and Angert 2015; Bauke et al. 2021).

To interpret the obtained isotope data, the following calculations are useful:

  • Isotopic equilibrium between O in phosphate and in water caused by the PPase (Chang and Blake 2015).

  • Effect of organic P hydrolysis via phosphomono- and -diesterases on δ(18O)PO4 values (see Table 5.1 under “Enzymes”).

The calculated isotopic equilibrium and the effect of organic P hydrolysis can then, for instance, be used in a mass balance to estimate the relative contribution of microbial P cycling and enzymatic hydrolysis of organic P to available soil P. Mass balances can also be used to estimate the contribution of different P pools, such as mineral P and inorganic P in plants to available soil P (Tamburini et al. 2012) (see also Chap. 3).

5.3.3 Application of the Method to Study Phosphorus Cycling in the Soil–Plant Continuum

With more protocols available for different P pools and environmental samples, the δ(18O)PO4 method is currently used to investigate a wider range of aspects of the P cycle in the soil–plant continuum (Table 5.3). See also Bauke (2021) for a more detailed discussion about soil δ(18O)PO4 values.

Table 5.3 Aspects of the phosphorus (P) cycle in the soil–plant continuum to which the δ(18O)PO4 method has been applied so far
Full size table

One of the first studies which used 18O in phosphate to study P cycling in soils was the study by Larsen et al. (1989). They added KH2PO4 labelled with 18O and 32P to soil, distributed the soil among different pots, and grew grass in those pots for three months. To exclude biotic activity, they also added germicide to some pots for three months. Half of the added 18O-label was lost in the untreated soils, whereas no loss was observed in the germicide-treated soil, which was attributed to a lack of biotic activity (for more information about labelling experiments see also Chap. 3). In line with this, Tamburini et al. (2012) showed by analysing δ(18O)PO4 values of soil and plant P pools along a soil chronosequence that prior to being released to the available soil P pool, phosphate is cycled through microbes. In an incubation experiment with organic horizons of two forest soils, differing in their amount of available P, Pistocchi et al. (2020) added 18O-labelled water to the soils. In the soil with a high P availability, δ(18O)PO4 values of available P approached equilibrium values, whereas in the low-P soil, the impact of hydrolyzing enzymes was visible in δ(18O)PO4 values of available P, thus showing that this pool was mostly replenished by P derived from organic P mineralization.

Whereas more and more studies use δ(18O)PO4 values to investigate P cycling in soils, less is known about plants, i.e. P acquisition by plants and plant internal P cycling, and how these processes affect δ(18O)PO4 values. Plants play a vital role for P cycling in soils, as they not only take up P from the soil solution, but also exude enzymes and create biopores thereby altering chemical and physical soil properties. Plant litter can also be a major source of P in soils (Sayer et al. 2020); therefore, it is necessary to understand how plants, under different environmental conditions, alter δ(18O)PO4 values. In a greenhouse experiment with soybeans, Pfahler et al. (2013) showed that δ(18O)PO4 values of inorganic P, extracted from plant leaves, are more enriched in 18O compared to P supplied in the nutrient solution. This was most likely caused by the O exchange between phosphate and leaf water, which is usually enriched in 18O compared to soil water, due to transpiration. In a follow-up study, Pfahler et al. (2017) investigated if δ(18O)PO4 values can be used to study plant response to P limitation. Combining δ(18O)PO4 values with radioactive 33P labelling, they showed that δ(18O)PO4 values of trichloroacetic acid (TCA)-extractable P from soybean leaves, subjected to P limitation, are less enriched in 18O compared to the soybeans growing with an ample amount of P in the nutrient solution. This was most likely caused by hydrolyzing enzymes like acid phosphatase, releasing, and then translocating P from senescent to younger leaves. Recently, several studies have investigated whether the δ(18O)PO4 method can be used to trace P inside plants after uptake (Qin et al. 2018; Hauenstein et al. 2020; Bauke et al. 2021). Qin et al. (2018) used 18O-labelled phosphate to trace applied phosphate from the soil to maize shoots. They found this possible only when the roots of maize plants were not inoculated with an arbuscular mycorrhizal fungus (AMF). They concluded that the 18O-label was lost during phosphate metabolism inside the AMF (Qin et al. 2018). Hauenstein et al. (2020) and Bauke et al. (2021) used 18O-labelled water instead of 18O-labelled phosphate to investigate P cycling in plants. The 18O-labelled water was provided to the plants as irrigation water or in the nutrient solution. Hauenstein et al. (2020) used this method to study P nutrition of beech at two forest fertilization experiments in Germany. Based on the analysis of δ(18O)PO4 values of inorganic P in the xylem sap, they postulate that in the P fertilized treatments, the additional P in the xylem was derived from fertilizer P and biologically cycled soil P. Bauke et al. (2021) used spring wheat as a model plant and conducted three different experiments using either hydroponic systems or pots filled with soil. Unlike shoots, root δ(18O)PO4 values of TCA P partly preserved the δ(18O)PO4 value of the P source due to lower metabolic activity in the roots. They therefore suggested that root P might provide better information on the P source than above-ground biomass, although their study did not consider the effects of mycorrhization, as in Qin et al. (2018).

As shown in the above-mentioned studies, plants can greatly alter δ(18O)PO4 values of assimilated phosphate. Hacker et al. (2019) showed that plant diversity, via its effect on evaporation and thus soil water δ(18O)H2O values, can also indirectly impact δ(18O)PO4 values of bioavailable soil P. To show this, two waters, differing in their δ(18O)PO4 values, were applied to 27 plots in the Jena experiment, which consisted of around 80 plots with different plant species combinations. Microbial P turnover influenced δ(18O)PO4 values of bioavailable soil P more in plots with high plant diversity compared to plots with lower plant diversity (Hacker et al. 2019).

Plants can potentially affect soil δ(18O)PO4 values in many ways, not just via their impact on evaporation. Plants, like microorganisms, exude enzymes that are known to alter δ(18O)PO4 values (see Table 5.1). However, little is known about how other plant acquisition strategies, such as symbiosis with AMF might affect soil δ(18O)PO4 values. To the best of our knowledge, only Wang et al. (2016) investigated the impact of low-molecular-weight organic acids (LMWOAs) on δ(18O)PO4 values. They used hydroxyapatite and different LMWOAs (acetic, oxalic, and citric acid) in batch and column experiments. The observed fractionation factors were relatively small: −0.3 to 1.1 ‰ (batch experiment) and −1.3 to 1.1 ‰ (column experiment).

5.3.4 Natural Abundance Versus Labelling Experiments

Most field studies work with the natural abundance of 18O in the environment. However, the differences between δ(18O)PO4 values of P pools between treatments or between different P pools are sometimes within 1 or 2 ‰ and it might be difficult to draw sound conclusions, especially when replication is missing (Pfahler et al. 2020b). One possibility is to work with strong environmental gradients, as in the pioneering field study by Tamburini et al. (2012) at a soil chronosequence. In other cases, the addition of an 18O-labelled P source or 18O-labelled water is advisable. Labelling experiments have great potential to bring further insights into the P cycle in the soil–plant continuum. Following the work by Larsen et al. (1989) in soils, Melby and co-authors conducted several studies using 18O-labelled phosphate (Melby et al. 2011, 2013a, b). In the first study, Melby et al. (2011) described a method how to produce 18O-labelled phosphate to be used in other studies. They then conducted two studies in which they used 18O-labelled phosphate to investigate biological P cycling (Melby et al. 2013b) and phosphate sorption (Melby et al. 2013a). In recent years, more and more studies, including the ones mentioned before, have applied an 18O-label to soils/plants, either via phosphate or via water. Most of them are conducted under controlled environmental conditions (i.e., greenhouse or growth chambers), since the application of a labelled compound in the open environment requires more care and might not always be suitable. In areas with high precipitation like tropical rainforests or high evaporation like deserts, 18O-labelled water applied to the soil might rapidly be altered, which needs to be accounted for when interpreting the obtained data (Kendall and McDonnell 2012; Beyer and Penna 2021). In addition to adding a 18O-labelled compound, also the addition of other labels like 33P/32P (Pfahler et al. 2017; Siegenthaler et al. 2020) and 13C can bring useful insights into P cycling (Gross and Angert 2017). In an incubation study, Siegenthaler et al. (2020) applied 32P as well as 18O-labelled water to soil samples from a climatic gradient in Hawaii. They found that both techniques complement each other like in case of NaOH-EDTA P where 32P and δ(18O)PO4 values both indicated that NaOH-EDTA P was only partly cycled during the incubation experiment. Combining δ(18O)PO4 with labelling an organic P compound with 13C, Gross and Angert (2017) showed that this combination could help elucidating both, C and P, cycling in the environment.

5.4 Challenges of the δ(18O)PO4 Method Application

The major pitfalls of the δ(18O)PO4 method can be grouped into three categories: sample handling mistakes, issues during purification, and lack of additional data for results interpretation.

5.4.1 Sample Handling Mistakes

δ(18O)PO4 is strongly influenced by biotic processes. It is therefore essential to reduce biological activity immediately after sampling to not alter the δ(18O)PO4 of any targeted P pool or compound. Drying soil samples might be effective in reducing biological activity, but drying and re-wetting are known to lyse microbial cells and release microbial P into the available P pool (Dinh et al. 2017). Thus, analysing δ(18O)PO4 of available P extracted from dried soils is different from δ(18O)PO4 values determined using fresh soils. A relatively simple way to reduce biological activity is to store the samples on ice in a freezer/polystyrene box. However, prolonged storage of fresh soil samples, even at 4 °C, might affect δ(18O)PO4 values, especially those of more labile P pools. Storing fresh soil samples for a minimum amount of time, even in the fridge, is thus recommended.

5.4.2 Extraction Issues

One issue when working with plants and δ(18O)PO4 is that enzymatic activity needs to be stopped during extraction, otherwise δ(18O)PO4 could be altered (Bieleski 1964). The combination of low temperatures and trichloroacetic acid was relatively efficient in stopping enzymes during extractions. Similarly, resin and microbial P extractions for δ(18O)PO4 analysis are usually conducted at 4 °C (Tamburini et al. 2012).

Another issue associated with extracting P for δ(18O)PO4 analysis from soil samples is that sometimes large soil volumes are required to obtain a sufficient amount of P (10–20 µmol P) for the purification protocol (Pfahler et al. 2020a). This is challenging for two reasons: (1) keeping the original soil-to-solution ratio for an extraction might be difficult and (2) reducing large volumes of extracts is laborious and needs to be done carefully (Adu-Gyamfi and Pfahler 2022). Keeping the original soil-to-solution ratio is important, as changes to it might result in extracting more or less P compared to the original ratio (Weiner et al. 2011; Pfahler et al. 2020a) and thus comparing results between studies is difficult. Large volumes of extracts need to be concentrated, for instance via brucite precipitation to approximately 100 mL to purify it and precipitate silver phosphate as the final analyte. Several brucite precipitations might be necessary, which could result in a loss of P if not done carefully (Adu-Gyamfi and Pfahler 2022).

5.4.3 Issues During the Purification Protocol

Adu-Gyamfi and Pfahler (2022) addressed the issues during the purification protocol in detail. In brief, issues during the purification of extracts are often due to high quantities of organic matter in the extracts or low P concentrations. There are ways to deal with those issues, but one must be aware of them first. An additional step with a DAX resin or (several) brucite precipitations could be included to remove more organic matter (Adu-Gyamfi and Pfahler 2021). Chemicals used in the purification protocol could potentially contain small impurities of P. Also, some detergents used in laboratories to clean equipment contain P. Thoroughly checking for P contaminations is therefore a good practice, especially when handling samples with low P concentrations.

Soils and soil extracts rich in calcium, iron, and silica might cause problems during the purification protocol (Adu-Gyamfi and Pfahler 2022; Tamburini et al. 2018). Knowledge of the chemistry of the studied soils is thus recommended to be able to prevent and face these problems.

5.4.4 Pitfalls in 18O-labelling Experiments

In addition to the above-mentioned pitfalls, there are additional ones when setting up labelling experiments with 18O-labelled water or a 18O-labelled P compound/pool. One pitfall might be the sensitivity and linearity of isotope ratio mass spectrometry (IRMS) measurements, as observed in the case of other isotope studies (Blessing et al. 2008). Measuring 18O-labelled and unlabelled samples in separate IRMS runs is advisable to avoid carry-over memory effects. Most available silver phosphate reference materials (store-bought) do not cover a wide range of δ(18O)PO4 values. Typically, they range between 14 and 23 ‰. It might therefore be necessary to produce an in-house standard (Lecuyer et al. 1999), covering the expected range of δ(18O)PO4 in the experiment (Halas et al. 2011; Watzinger et al. 2021).

Like for other labelling experiments, background information of the soil- or plant-inherent P concentration and isotope values is necessary to allow choosing an adequate labelling strength. Especially in samples with high P concentrations, the added label may be strongly diluted, and the intended processes will not become apparent. It is recommended to perform a preliminary test consisting of four steps: (1) analysis of δ(18O)H2O of soil water, (2) analysis of δ(18O)PO4 in different soil P pools, (3) calculate theoretical equilibrium δ(18O)PO4 values using different values for δ18O of soil water, and (4) choose a δ(18O)H2O of soil water which would result in a well distinct, for instance, by a factor of 5 or 6, δ(18O)PO4 equilibrium value. For label application, a homogenous distribution of the labelled substance (water or P compound/pool) to a soil (in the field or pot) is crucial for 18O-labelling experiments. Sample homogeneity could also be an issue when using 18O-labelled water or P compounds. During the experiment and subsequent purification or sample treatment procedures, labelled samples should also be handled with care and contact with unlabelled samples should be avoided to avoid any carry-over of the label.

5.4.5 Lack of Additional Information

Phosphate mainly exchanges O with water, and δ(18O)H2O of water plays an important role in interpreting δ(18O)PO4 values. Analysing δ(18O)H2O of water in soils and plants requires the extraction of the water, usually via cryogenic distillation. For more information about cryogenic distillation and discussions about it see, for example, Orlowski et al. (2016) and references therein. Evapotranspiration influences δ(18O)H2O of soil and plant water. Putting sub-samples of the soil and plant samples into appropriate containers directly at the time of sampling is therefore recommended. Because the equilibrium promoted by PPase is temperature-dependent, measuring the ambient/soil/leaf temperature is highly recommended. One of the major aspects of P cycling in the soil–plant continuum is the P availability. Determining at least the concentrations of available P (soils), total P (soils, plants), and inorganic P (plants) is thus recommended. Knowledge of other soil properties, such as soil pH, organic matter content, and bulk density, might also help in interpreting δ(18O)PO4 values.

5.5 Way Forward and Prospects of the δ(18O)PO4 Method for Plant P Cycling

5.5.1 Earth System Models and δ(18O)PO4

Recently, P has been incorporated into Earth system models in addition to carbon (C) and nitrogen (N) (Reed et al. 2015). Advances in those models could be made if we have a better understanding of P mineralization, how P could limit C cycle processes, P sorption, and the coupling between P, C, and N cycling (Reed et al. 2015). δ(18O)PO4 has the potential to advance our understanding in these aspects, especially when biotic processes are involved. Studies conducted so far indicate that δ(18O)PO4 could provide useful information about the role of soil pH, enzymes, and soil microorganisms for P cycling in the soil–plant continuum, the effect of P limitation, the interaction of P with other nutrients such as N, and the residence times of P in different soil P pools (see Sect. 5.3). The first attempts to use δ(18O)PO4 for modelling/estimating P fluxes exist (Tamburini et al. 2012; Jaisi et al. 2017; Pfahler et al. 2017; Helfenstein et al. 2018; Bauke et al. 2021), but further research is needed.

5.5.2 δ(18O)PO4 Values of Single Organic Phosphorus Compounds

To advance Earth system models even further, knowing δ(18O)PO4 values of specific organic P compounds might be useful. Knowing those values, one could more accurately estimate the parameters associated with P mineralization. The determination of δ(18O)PO4 of organic P usually requires treating the pure organic P compound with UV light to hydrolyze the organic P and release inorganic P (Blake 2005; Tamburini et al. 2018). Wu et al. (2015) and Sun et al. (2017) determined δ(18O)PO4 of phytate via direct pyrolysis in the TC/EA-IRMS. So far, studies which analysed δ(18O)PO4 values of single organic P compounds have always used pure compounds and did not extract them from environmental samples. The next step might be to investigate whether some of the available methods for extracting single organic P compounds from environmental samples are suitable for the determination of δ(18O)PO4 values. Phytate can be isolated from soil samples via hypobromite oxidation (Turner 2020). However, due to the experimental conditions, it is necessary first to adapt this method to the phosphate purification protocols currently available, so that reliable soil phytate δ(18O)PO4 values can be obtained.

5.5.3 More Laboratory and Glasshouse Studies Needed to Advance the Application of δ(18O)PO4 in the Field

So far δ(18O)PO4 studies have focused on hydrolyzing enzymes but data on the fractionation factors of synthesizing enzymes and phosphotransferases, to name a few, for the δ(18O)PO4 are missing. Some papers, however, report isotopic effects of 18O on the rate constants of these enzymes, for example Hengge et al. (2002). Information about those enzymes is vital, especially for the P cycle in microorganisms and plants. How microorganisms deal with low soil pH and low soil P conditions simultaneously is also not well understood (Lidbury et al. 2017). The δ(18O)PO4 might help shed light on this, as indicated in the study of Pfahler et al. (2020b). Glasshouse studies excluding variables such as changes in the vegetation but focusing mainly on the soil pH might therefore be useful. As mentioned by Reed et al. (2015), better knowledge about P acquisition strategies is necessary to improve models. Excreting organic acids is one P acquisition strategy, and one study showed that the dissolution of hydroxyapatite by organic acids (oxalic, citric, and acetic acid) leads to a small isotopic fractionation (Wang et al. 2016). As organic acids mainly affect processes in the rhizosphere, studies determining δ(18O)PO4 values in the rhizosphere are encouraged. This is challenging due to the small amounts of rhizosphere soil that can be collected. Rodionov et al. (2020) found no differences in δ(18O)PO4 values of HCl-extractable P between bulk and rhizosphere soil samples. δ(18O)PO4 values of the more actively cycled P pools, such as microbial and available P, extracted from rhizosphere samples have not yet been reported.

5.5.4 Lack of Field Studies Including Plants and Ecosystems

In general, information about δ(18O)PO4 of plants is still limited (see Sect. 5.3). Detailed plant studies are thus far only available for soybeans (Pfahler et al. 2013, 2017), maize (Qin et al. 2018), and spring wheat (Bauke et al. 2021). Additionally, Tamburini et al. (2012) analysed different plant species along a soil chronosequence at the Damma Glacier in Switzerland, Helfenstein et al. (2018) reported δ(18O)PO4 values for different plant species at a climatic gradient in Hawaii, and Pfahler et al. (2020b) determined δ(18O)PO4 of bulk vegetation samples from a grassland experiment in the UK. Those studies provided useful hints/insights about using δ(18O)PO4 in plant studies, but data is still limited. The effect of P limitation has so far only been investigated in soybeans, and the study by Pfahler et al. (2020b) indicates that P limitation has a similar effect on grasslands. Studies about the effect of P limitation on major crops, such as maize, rice, and wheat, especially under field conditions, are still missing. Likewise, there are no studies using δ(18O)PO4 and investigating the effect of crop rotation or intercropping on P cycling. Crop rotations or intercropping often combine plants with different P acquisition strategies to increase the P use efficiency of crops, and it could be expected to have an influence on the δ(18O)PO4 of soil and plants.