Role of magnesium fertilisers in agriculture: plant–soil continuum

Additional keywords: antagonism, mineral fertilisers, soil texture, solubility.

Global human population has doubled during the last 45 years and this trend will continue in the coming decades. The availability of adequate potable water and water reserves for agricultural purposes is already critical in many regions, regardless of change in global climate. In the very near future, a strong increase in crop production will be needed to meet food and energy demands alongside preservation of the ecological and energy-related resources of our planet. One of the major challenges for agriculture will be to maintain the crop yields in harsher environments (e.g. seasonal drought, heat and excess light energy) and/or enhance crop yields in more resource-efficient systems (Reynolds et al. 2012). Minimising the ‘yield gap’ and increasing yield stability under different stress conditions are of strategic importance to guarantee food for the future (Cattivelli et al. 2008). Recent research advances in crop physiology and genomics have led to new insights into stress tolerance and provided breeders with new knowledge and tools for plant improvement (Tuberosa and Salvi 2006). In parallel, innovative site-specific agricultural management techniques and precise crop-nutrient management also need further attention so that crop plants with higher yield potential can be grown under unfavourable conditions and can achieve their potential yield under given environmental conditions. Better understanding of crop strategies to deal with stress situations and of factors that control response mechanisms is key to developing innovative techniques.

For successful farming practice and optimal crop-nutrient management, fundamental principles of mineral plant nutrition must be considered, such as physical, chemical and biological processes in plants and soils. The International Plant Nutrition Institute (IPNI), which performs agricultural projects in collaboration with scientists from various countries, aims to develop a simple management concept (BMP, best management practise) to enhance specific crop performance under certain environmental conditions. The 4R Nutrient Stewardship technology initiated by IPNI is one of the most advanced tools for decision making with site- and crop-specific application. There is direct interaction between applying the right nutrient source, at the right rate, right time and right place (4R), and profitable impacts for good crop growing, soil health and decreased pollution of the environment via enhanced nutrient-use efficiency.

Magnesium (Mg) is involved in many physiological and biochemical processes; it is an essential element for plant growth and development and plays a key role in plant defence mechanisms in abiotic stress situations (Cakmak and Kirkby 2008; Cakmak and Yazici 2010; Cakmak 2013; Gransee and Führs 2013; Huber and Jones 2013; Mengutay et al. 2013). The most commonly known function of Mg in plants is probably its role as the central atom of the chlorophyll molecule in the light-absorbing complex of chloroplasts and its contribution to photosynthetic fixation of carbon dioxide (Cakmak and Kirkby 2008; Cakmak and Yazici 2010; Gerendás and Führs 2013). However, the Mg bond to chlorophyll makes up only a small part of the total Mg fraction. Depending on the Mg status of the plant, ~20% (Marschner 2012; Gransee and Führs 2013) and up to 35% (Cakmak and Kirkby 2008; Cakmak and Yazici 2010) of the element is localised in the chloroplast, and the remaining Mg is present in more mobile forms (Marschner 2012). Because of its high phloem mobility, Mg can easily be translocated to active growing parts of the plant where it is needed for chlorophyll formation, enzyme activation for protein biosynthesis, and phloem export of photosynthates to ensure vegetative and generative growth. Therefore, first visual deficiency symptoms generally occur on older leaves (Cakmak and Kirkby 2008; White and Broadley 2009; Gransee and Führs 2013). Thus, even slight Mg deficiency may affect biomass formation and plant susceptibility to environmental stresses by diminishing several biochemical and physiological processes.

Magnesium is a common constituent in many minerals, comprising 2% of Earth’s crust. However, most soil Mg (90–98%) is incorporated in the crystal lattice structure of minerals and, thus, not directly available for plant uptake. Owing to high variation in Mg content of source material and the degree of weathering, the total content of Mg in soils varies considerably, between 0.05% and 0.5% (Grimme 1991; Maguire and Cowan 2002; Gransee and Führs 2013). Many common soil minerals contain Mg, including amphibole, biotite, chlorite, dolomite, montmorillonite, olivine, pyroxene, serpentine and vermiculite. Stores of bioavailable Mg originate from inputs by mineral weathering. Therefore, soils that have developed from coarse-grained rocks low in these minerals tend to be low in Mg.

Soil Mg is often subdivided into four fractions: rapidly exchangeable, slowly exchangeable (acid-soluble), organic complexed, and structural forms (Mayland and Wilkinson 1989) (Fig. 1). The last of these accounts for differences in bioavailability; plants absorb Mg from the soil solution, which is buffered by the readily exchangeable form, which, in turn, is slowly replenished by the soil reserves. Soil texture is a key variable that affects plant-available Mg. Because Mg is located in clay minerals and associated with cation exchange sites on clay surfaces, clayey soils generally contain adequate Mg for plant requirements, whereas sandy soils are frequently deficient in Mg (Mayland and Wilkinson 1989). Several ferromagnesian minerals (e.g. olivine, pyroxene, amphibole, and mica) are major Mg sources in basic igneous rocks (Chu and Johnson 1985). Secondary minerals such as magnesite, talc, and the serpentine group are the weathering products of these primary minerals. Salmon (1963) suggested that the main sources of Mg that can be made available in soils are secondary minerals, particularly clay minerals, mica, and chlorite. To become soluble, Mg adsorbed on a clay particle needs to be replaced by other cations, such as potassium (K⁺) and hydrogen (H⁺), from the soil solution. Stahlberg (1960) determined the amount of slowly exchangeable Mg released from several Swedish topsoils (boiling them in 1 n HCl) and summarised that vermiculite and chlorite were the main sources of this acid-soluble Mg. However, there was a poor correlation between exchangeable Mg and total or acid-soluble Mg in soils (Prince et al. 1947; Stahlberg 1960; Baker 1971). Similarly, Hailes et al. (1997a) observed that exchangeable Mg was not significantly (P > 0.05) correlated with organic carbon, and only 45% of the variation in exchangeable Mg could be explained by a combination of pH and clay content. Therefore, plant-available Mg concentrations cannot be accurately predicted solely on the basis of the parent material composition, because of differences in mineral weathering rates depending on various environmental conditions. Duration and intensity of weathering, soil moisture, soil pH, and root and microbial activity in soil are other key factors that determine the plant-available Mg release from soils (Mayland and Wilkinson 1989).

Fig. 1.  Interrelationship of magnesium forms in the soil and the effects of soil microorganisms and root exudates (after Zörb et al. 2014).

Soil pH has a direct effect on release of Mg from clay minerals, as well as on plant Mg uptake. Chan et al. (1979) and Hailes et al. (1997b) showed that Mg that is still exchangeable at a soil pH <6.0 becomes non-exchangeable when soil pH is increased to >6.5. Similarly, Sumner et al. (1978) reported that when the pH of Ultisols increased from 5.5 to 7.5, soil exchangeable Mg dropped >50%. Because plant uptake of Mg can be hampered by an excess concentration of other cations (e.g. H⁺), higher soluble Mg concentrations in soil solution do not necessarily mean that this Mg is plant-available (Metson 1974). In acid soils (pH <5), high levels of exchangeable aluminium, which is harmful for plants, are also released. Additionally, the soil solution is saturated with H⁺ ions instead of base cations at the site of rhizosphere (Metson 1974). In this context, the decrease in plant availability of Mg at low soil pH is a consequence of the increasing inability to build up and maintain a sufficient pH and hence electrochemical gradient across the plasma membrane of root cells (Schubert et al. 1990; Gransee and Führs 2013). In conclusion, at low pH, soil exchangeable Mg concentration may increase; however, the dominance of H⁺ at the site of rhizosphere may interfere with the uptake of Mg, causing Mg deficiency and hampering yield and quality of agricultural products (Mayland and Wilkinson 1989).

Magnesium is well described for its important role in chlorophyll synthesis. In addition, Mg plays a key role in several plant physiological processes through its key function in phloem loading, being a co-factor and allosteric modulator for >300 enzymes (including Calvin cycle, kinases, RNA polymerases and ATPases), and in chelation to nucleotidyl phosphate forms (Cowan 2002; Shaul 2002; Verbruggen and Hermans 2013). Therefore, Mg is crucial for the transport of assimilates from source leaves to sink organs, and thus, an early symptom of Mg-deficiency stress in plants is the disturbed partitioning of assimilates between roots and shoots, resulting in increased accumulation of these assimilates in source leaves and reduced growth rate of sink organs (Cakmak and Kirkby 2008; Cakmak 2013). Here, the root system and other developing plant parts (young leaves and grain) as a heavy sink for assimilates suffer from the impaired phloem loading. The limited availability of carbohydrates and other important assimilates leads to a reduction in root growth, promoting the risk of water and nutrient deficiencies from reduced exploration of soil volume, causing less access to soil resources (Cakmak and Kirkby 2008; Cakmak 2013; Gransee and Führs 2013).

Disturbance in carbon partitioning may be regarded as a latent deficiency symptom (Gransee and Führs 2013) and it occurs long before visible symptoms such as interveinal chlorosis (e.g. in wheat or maize; Cakmak and Yazici 2010). Because of impaired phloem loading, accumulation of carbohydrates in leaves suffering Mg deficiency often causes a decrease in CO₂ fixation by Ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco), in very early stages of the deficiency. This occurs by two main mechanisms: (i) feedback inhibition of sucrose synthesis; and (ii) accumulation of starch in the chloroplast, affecting CO₂ conductance of the chloroplast membrane and causing lower CO₂ partial pressure at the catalytic site of Rubisco (Araya et al. 2006). Starch synthesis and accumulation in the chloroplasts is known to cause deformation of the chloroplasts and a decrease in the rate of CO₂ diffusion from the membrane (Nafziger and Koller 1976; Keenan et al. 2010). Thus, imbalance between light capture and its utilisation typically occurs, in which the non-utilisation triggers the production of reactive oxygen species (ROS) in Mg-deficient plants (Cakmak and Yazici 2010). Whether enhanced ROS would serve as signalling molecules and/or could cause oxidative damage to the chlorophyll molecules depends on the delicate equilibrium between ROS production, and their scavenging (Asada and Takahashi 1987; Elstner 1991; Mittler 2002). As shown by some examples in various crops (Fig. 2), one of the early visual symptoms of leaves suffering Mg deficiency is interveinal chlorosis, meaning chlorophyll degradation due to excessive ROS production (Mengutay et al. 2013). In almost all cases, the reason for such chlorosis in Mg-deficient leaves is excessive ROS damage, not lack of Mg for chlorophyll synthesis.

Fig. 2.  Magnesium deficiency symptoms in crops. Arable crops: A, barley (Hordeum vulgare); B, wheat (Triticum aestivum L.); C, rye (Secale cereale); D, maize (Zea mays L.); E, maize Mg-deficient (left), adequately supplied (right); F, soybean (Glycine max L.); G, oilseed rape (Brassica napus); H, potato (Solanum tuberosum); I, sugar beet (Beta vulgaris L.). Fruit: J, wine (Vitis vinifera); I, banana (Musa L.); J, raspberry (Rubus idaeus); M, apple (Malus domestica). Vegetables: N, cucumber (Cucumis sativus L.); O, bell pepper (Capsicum annuum); P, pea (Pisum sativum); Q, tomato (Solanum lycopersicum), Mg-deficient (left), adequately supplied (right). Cash crops: R, tobacco (Nicotiana tabacum); S, cotton (Gossypium hirsutum); T-U, coffee (Coffea arabica). (Sources: A–E, G–M, O–Q, K+S KALI GmbH; T–U, IAPN; F, IPNI, EAB Francisco; N, IPNI, AC Tellez Andrade; R, IPNI, B Zhu; S, IPNI, ChS Rao).

Reducing the environmental issues while increasing unit productivity of nitrogen (N) fertiliser is the aim in modern agricultural practice (Grzebisz et al. 2010). Vegetative and generative plant growth greatly depend on the plant’s ability to take up N in amounts necessary to cover metabolic requirements at every stage of its lifecycle (Andrieu et al. 1997; Grzebisz 2013). Plant access to N and/or its utilisation can be negatively affected by physiological disorders of plants under other nutrient deficiencies (e.g. Mg deficiency). Because Mg is involved in simultaneously controlling processes responsible for photosynthesis, assimilate production and partitioning among plant parts, it seems to be a major player in N uptake and its utilisation (Gastal and Lemaire 2002; Shaul 2002; Rubio et al. 2003; Cakmak and Kirkby 2008; Grzebisz 2013). The review by Grzebisz (2013) compared various crops (sugar beets, cereals and maize) for their ability to take up more Mg in deficient soils and estimated its effect on N-use efficiency (NUE) and yield components. For example, Grzebisz et al. (2010) reported that, compared with other crops tested, tuber and root crops (especially sugar beet) showed the best response to Mg fertiliser supply, with the greatest increase in crop yield from enhanced NUE (Grzebisz et al. 2010). In N-limited environments, sufficient Mg supply was also reported to enhance N uptake (soil + fertiliser) (Grzebisz et al. 2010; Grzebisz 2013). This is due to: (i) the importance of Mg in assimilate translocation from source to sink organs, which increases root growth and may enhance crop–microbe competition for N uptake (Casson and Lindsey 2003; Cakmak and Kirkby 2008); and/or (ii) enhanced NUE from better transport of amino acids and increased rate of photosynthesis (Andrieu et al. 1997; Grzebisz et al. 2010).

In the field, crops face multiple stress factors that may negatively affect metabolism, growth, and yield formation of the cultivated plants—abiotic stresses (e.g. drought, salinity, cold or high light events) and biotic stresses. Furthermore, climate models predict that incidences and duration of drought and heat-stress events will increase in some parts of the world, and indicate that such conditions will have a dramatic impact on agricultural production and farming practices in coming decades (Brouder and Volenec 2008). Because of its fundamental roles in plants, Mg nutrition affects the resistance of crops to most abiotic and biotic stresses, both directly and indirectly.

Under drought conditions, a common consequence is limited nutrient uptake (including Mg) from the soil, because in dry soils, diffusion and mass flow of nutrients to the root are hampered (Engels and Kirkby 2001). Additionally, root growth is inhibited in dry soils and this may lead to a further reduction in nutrient uptake. Plant roots may counteract this physical problem by increasing root hair length and enhancing the secretion of a gelatinous substance called ‘mucilage’ with high water-holding capacity (Carminati and Vetterlein 2013). As discussed above, suboptimal Mg nutrition affects phloem loading, causing an accumulation of sucrose in photosynthetically active tissues and a poor energy supply of roots. Therefore, we may speculate that plants suffering from Mg deficiency will be more sensitive to soil-water deficit, because root exudation (e.g. mucilage) and root elongation require enormous assimilate transport from the source organs (e.g. photosynthesising leaves).

Under Mg deficiency, sugar accumulation in the leaves causes feedback inhibition of Rubisco and lowers the rate of photosynthesis. The light reaction is also dependent on Mg providing charge balance. Inhibition of photosynthesis by drought or Mg deficiency results in a misallocation of electrons to oxygen, thereby producing ROS, which cause oxidative stress (Cakmak 2005). Heat stress, high light intensity and atmospheric drought often co-occur with soil-water deficit (Braun et al. 1996). Similar to Mg deficiency, heat stress causes peroxidative damage in chloroplasts, excessive ROS production and enhanced antioxidative defence enzymes (Gong et al. 1997; Dash and Mohanty 2002; Jiang and Huang 2002). Harmful effects of excess heat stress on arable crops are pronounced when plants are simultaneously exposed to low Mg supply, and adequate Mg nutrition is critical in plant response to heat-stress events (Mengutay et al. 2013). Therefore, to maintain growth and yield formation, the amount of fertiliser needed to meet the crop Mg requirement is higher under drought than under well-watered conditions. Selecting plant genotypes with higher potential to take up more Mg and with higher Mg-utilisation efficiency should be considered as an option to secure optimum yield under stress conditions.

Application of foliar Mg fertilisers is also reported to ameliorate the nutritional status of crops subjected to Mg deficiency. It has been demonstrated to increase the chlorophyll concentration and vegetative yield of plants (Neuhaus et al. 2014). However, little is known about the plant’s ability to take up nutrients from the leaves treated with foliar fertilisers (Eichert et al. 2008). More research is needed to reveal the leaf Mg uptake potential of various crops or genotypes. Under drought situations, nutrient uptake from the leaves is limited because of closed stomata or changes in leaf morphology (e.g. thicker leaves and waxy leaf surface). In agricultural systems with higher risk of drought periods, application of Mg fertiliser, in particular in combination with other nutrients, is suggested to enhance fertiliser-use efficiency (Römheld and Kirkby 2010).

Once the Mg ion has reached the root surface, it can be taken up into the root cells by building up an electrochemical gradient through pumping protons out of the cytoplasm, which allows the passive influx of Mg into the root cells (Barber 1984; Mayland and Wilkinson 1989). However, especially in sandy soils, application of high rates of K or ammonium (NH₄⁺) fertiliser often enhances the risk of Mg deficiency (Mulder 1956). High concentrations of these cations in the soil solution interfere with Mg uptake by plants (called nutrient antagonism). Usually, such events do not occur when the soil contains more exchangeable Mg than exchangeable K (Mulder 1956; Metson 1974; Seggewiss and Jungk 1988). The simplified drawing in Fig. 3 shows how antagonism between Mg and K possibly occurs in soils containing various concentrations of Mg and K. In the case of a higher Mg concentration in the soil solution, generally K uptake is not disturbed (Fig. 3c). Typically, the amount of K in the soil solution is much lower than Mg concentration; therefore, plants have been developed specific K-transport systems in the root cells to ensure sufficient K uptake when its concentration in the soil solution is critically low (Horie et al. 2011). These specific K transporters cannot be blocked by other nutrients. By contrast, the Mg transporters are non-specific and can be passed by other cations such as K. Therefore, when K concentration in the soil–root interface is high, plant ability to take up sufficient Mg is limited. Seggewiss and Jungk (1988) and Wilkinson et al. (1990) reported that Mg uptake at the root surface was inhibited by K concentrations >20 µmol L^–1. Additionally, ratios of Mg accumulated in shoots to whole plant Mg content are reportedly negatively correlated with plant K concentrations in roots (Huang et al. 1990). This suggests that excessive K concentration in roots may depress the rate of net Mg translocation from roots to shoots. However, further studies are needed to confirm this hypothesis.

Fig. 3.  Antagonistic effect of potassium (K) and magnesium (Mg): model for nutrient uptake from the soil solution (after Marschner 2012). Left of arrows: nutrients in the soil; right of arrows: nutrients in the root.

As a competing ion in cation exchange reactions, calcium (Ca) in excess may also interfere (similar to K) with Mg uptake (Metson 1974; Mayland and Wilkinson 1989; Wilkinson et al. 1990). This does not take place at low concentrations of Ca in the soil solution (Fageria 1973, 2009). Although Mg has a higher mobility in the soil than Ca (van der Heijden et al. 2013), Mg uptake has been repeatedly shown to be lower in soils with high Ca concentration in the soil solution (Ferguson and Clarkson 1976; Wilkinson et al. 1990; Diem and Godbold 1993; Gransee and Führs 2013). Similar to K and Ca antagonisms with Mg, NH₄ nutrition has been shown to decrease Mg uptake through (i) its acidifying character when assimilated, and (ii) ion competition at the adsorbing surfaces of the roots (Mulder 1956). Therefore, in soils with limited plant-available Mg, use of NO₃-based fertilisers is recommended because they do not interfere with Mg uptake (Mulder 1956). Lasa et al. (2000) showed that NH₄ nutrition results in an inhibition of growth in sunflowers in a sandy soil; however, additional Mg supply mitigated the adverse effect of NH₄ supply. Therefore, we may summarise that the use of NH₄-based N fertilisers enhances crop Mg demands compared with plants supplied with NO₃-based fertilisers.

The amount of Mg released from soil minerals is mostly small compared with the amounts needed to sustain high crop yield and quality. Additionally, conditions such as high or low soil pH, drought and high levels of competing cations (e.g. K⁺, NH₄⁺ and Ca²⁺) may reduce plant availability of Mg even if its concentration in soil solution is high (Fig. 3). Therefore, on soils with limited plant-available Mg for optimal crop production, the application of Mg fertilisers is crucial. Soil analysis and specific crop requirements need to be considered for accurate fertiliser management. For example, critical soil test value for 90% relative yield was defined as 0.21 cmol₍₊₎ kg^–1 of exchangeable Mg or 7% Mg saturation in a glasshouse experiment, whereas the critical (90% yield) plant-tissue Mg concentration (whole shoots) was 0.15% (Hailes et al. 1997b). However, further experimental work is needed to define critical soil test values for specific crops under field conditions.

Common Mg fertilisers are generally distinguished into two classes: soluble sources and semi-soluble sources. Minerals such as dolomite are semi-soluble Mg sources. These minerals are often used as Mg fertiliser with minimum processing. On the other hand, several naturally occurring soil minerals (such as kieserite) mined from the ground contain Mg in hydrated form (MgSO₄) and are used as soluble Mg fertilisers; these include magnesium sulfate monohydrate and/or magnesium sulfate heptahydrate (Kawamura and Rao 2007). The release rates of available Mg to soil solution from various mineral fertilisers are determined by their physical and chemical compositions, which are based on particle size and water solubility (Mayland and Wilkinson 1989; Härdter et al. 2004; Loganathan et al. 2005).

Water solubility of Mg in fertilisers mainly depends on its chemical composition, such as oxide, sulfate, carbonate, nitrate, chloride, phosphate or silicate (Mayland and Wilkinson 1989). Solubilities of various Mg-containing minerals are listed in Table 1. Minerals containing Mg in the form of MgSO₄ are completely soluble and thus most suitable for Mg-deficient soils. Alternatively, synthetic forms of MgSO₄ produced by a chemical reaction are also commercially available. Here, magnesium oxide reacts with sulfuric acid, producing a synthetic MgSO₄ (SMS) (Kawamura and Rao 2007). Purification, by separation from extraneous materials, and fine grinding (particularly for the less soluble sources) are the usual methods to prepare Mg-deriving sources (Metson 1974).

Table 1.  Solubility of magnesium minerals in water at 20°C
Struvite, dolomite and magnesite calculated from solubility products from respective source references. Sources are: 1, D’Ans and Lax (1949); 2, Seeger et al. (2011); 3, Bhuiyan et al. (2007); 4, Helgeson (1969); 5, Bénézeth et al. (2011)

The roles of Mg and the specific requirement for Mg supply differ with crop type. Below, we review three very different crops: grassland, wheat (Triticum aestivum L.) and potato (Solanum tuberosum L.).

Continued inputs of N, K and Ca fertilisers into grassland and the removal of soil Mg via leaching and transfer in animal products without sufficient resupply generally cause significant decline in available Mg in grassland soils. Therefore, Mg fertiliser applications to pastures are necessary to ensure balanced nutrient supply, not only to the plant (mixed-herbage pasture) but also to the animals. This is of great importance for livestock, because it reduces the risk of low pasture Mg intakes by grazing dairy and beef cows, which can lead to hypomagnesaemic grass tetany (hypomagnesaemia), a condition that usually occurs soon after calving. Hypomagnesaemia is a major cause of lower milk production, affecting ~30–50% of dairy herds in the major dairying areas of New Zealand (O’Connor et al. 1987; Loganathan et al. 2005). Edmeades (2004) cited data showing a significant decline in soil Mg levels in New Zealand pastoral soils from the 1980s to 2000.

Magnesium is important for both product quality and yield of cereals (Beringer and Forster 1981; Grzebisz 2013). Carbohydrate translocation and therefore optimal grain filling is positively supported by an optimal amount of available Mg. Thousand-grain weight, one of the most important wheat-grain quality parameters, is generally negatively affected in Mg-deficient soils (Marschner 2012; Grzebisz 2013). Processing behaviour is another grain-quality parameter (i.e. milling performance during flour production) known to be directly related to plant Mg content (Greffeuille et al. 2006; Gerendás and Führs 2013). In a field trial, wheat was supplied with MgSO₄ next to a basic NPK fertilisation, and it was clearly showed that Mg fertilisation increased grain yield, 1000-grain weight, crude protein and raw gluten in single grains (Al’shevskii and Derebon 1982). For Mg fertiliser management, specifically in sandy soils, Mg removal from grain and/or straw may need to be considered in order to keep adequate reserves of soil Mg.

Potato is a crop with a wide spectrum of quality parameters. Because of its usage in many forms (e.g. fresh consumption, processing and starch production), quality control is difficult and Mg is known to be a key element of potato quality (Talburt and Smith 1997; Hiltrop 1999). We can summarise the effect of Mg on potato yield and quality with four main points.

First, starch content and thus the mealiness grade of cooking potatoes vary because of enhanced assimilation and carbohydrate translocation (Smith 1977; Talburt and Smith 1997; Feltran et al. 2004). Therefore, the involvement of Mg in photosynthesis and assimilate translocation has a direct effect (Cepl 1994; Poberezny and Wszelaczynska 2011; Affleck et al. 2012).

Second, tuber firmness and resistance against mechanical stresses occurring during harvest, transport and storage are major quality parameters. Increased firmness, which is positively influenced by Mg supply, reduces the risk of bruising and various forms of discoloration (Klein et al. 1982).

Third, colouring of potatoes is most important for potato attractiveness. Discolorations of crude pulp and black spot incidence are induced by an enzymatic process, by harvest or transport, as well by processing of crude pulp. During processing, polyphenol oxidases in potato tubers interfere with free phenolic compounds. In this process, polyphenol oxidases catalyse the formation of diphenols from monophenols, changing finally to dark melanins (Mulder 1949; Mondy et al. 1965; Mondy and Koch 1978; Klein et al. 1981; Muneta 1981). The role of Mg in these reactions is not clear, but the role of Mg in assimilation and carbohydrate translocation indicates that an improved Mg supply should have a positive effect.

Fourth, another process is the production of glycoalkaloids, correlating with the greening of tubers (Maga and Fitzpatrick 1980). These alkaloids are described as toxic substances but the positive response of their concentration in leaves and stems lead to an increase of N supply (Love et al. 1994; Rogozinska and Wojdyla 1999), and the significance of Mg for N metabolism has prompted several investigators to study the influence of Mg supply on glycoalkaloid accumulation in potato tubers. Mondy and Ponnampalam (1985) and Klein et al. (1982) reported an increased N and protein concentration. Additionally, Evans and Mondy (1984) and Mondy and Ponnampalam (1985) described a significant increase in glycoalkaloid content after Mg application. Within these results, a hypothesis is raised that increased chlorophyll synthesis, as well as stimulation of sugar metabolism and/or changes in amino acid production, are part of this effect.

Unlike other cations, Mg is very mobile in soils because it is less bound to the soil charges. This results in a relatively high abundance of this element in the soil solution and thus a higher risk of leaching (Maguire and Cowan 2002; Shaul 2002; Gardner 2004; Gransee and Führs 2013). Leaching of Mg should be less severe in soils under a crop than under bare fallow but may increase when fertilisers are added. The potential for applied Mg to be taken up by crop plants and not lost via leaching greatly depends on the solubility of Mg fertilisers (Loganathan et al. 1999; Mitchell et al. 2000; Härdter et al. 2004). Applications of Mg fertilisers known as slow-release fertilisers (e.g. dolomite, magnesite and calcined magnesite) may mitigate leaching risks but not deliver sufficient plant-available Mg to crops. On the other hand, application of soluble Mg fertilisers (e.g. kieserite and SMS) may lead to Mg losses by leaching when applied to sandy soils with high water conductivity (e.g. sandy soils), specifically in wet seasons (Härdter et al. 2004; Loganathan et al. 2005).

In several studies, various Mg fertiliser sources were tested in different soils to examine Mg plant uptake and losses through leaching (Durrant and Draycott 1976; Heming and Hollis 1995; Härdter et al. 2004; Hanly et al. 2005). It is commonly accepted that there is almost no Mg leaching risk of slow-release Mg fertilisers (dolomite or fertilisers contains Mg in the form of Mg oxide). Efficiency of slow-released Mg fertilisers may be slightly higher, especially in acid soil conditions, and/or when applied in ground forms (Härdter 1992; Heming and Hollis 1995; Härdter et al. 2004). During the critical vegetative periods (such as shooting or flowering in wheat), crop nutrient requirements are at their maximum. Thus, it is important that applied fertilisers release sufficient plant-available nutrients during such critical periods when crop nutrient demand is high. In this context, water-soluble Mg fertilisers (such as kieserite and SMS) typically release more plant-available Mg in a relatively short period and are therefore more effective to secure crop Mg demand in Mg-deficient soils. On the other hand, SMS may cause higher risk of Mg leaching under certain conditions, such as in sandy soils and after heavy rain showers (Durrant and Draycott 1976; Heming and Hollis 1995; Härdter et al. 2004; Hanly et al. 2005; Loganathan et al. 2005). Härdter et al. (2004) reported that maize grown in a pot experiments had 19.6% greater Mg uptake and 10.6% higher yield with kieserite (water-soluble) treatment than MgO (slow-release Mg source) in both sandy and loamy soils. However, the same authors concluded that in soils with water-soluble Mg fertilisers applied, 9% and 22% of applied Mg could be lost via leaching during heavy rain showers.

Magnesium leaching of various water-soluble fertilisers may also vary depending on their chemical composition, granule size, and effect of soil pH and cation balance. Härdter et al. (2004) compared SMS and kieserite at two application rates and found that Mg leaching was ~33% lower in soils treated with kieserite than SMS, which they attributed to slower dissolution properties of kieserite than SMS. Slower release of Mg in kieserite may lead to a greater Mg adsorption in soil and thus lower leaching potential.

A similar experiment was done in another incubation trial at the Institute of Applied Plant Nutrition, Göttingen, where sandy soil with low soil pH was placed in small lysimeters. Briefly, after soil sampling (from Ahlten, in Hannover, Germany), soil was air-dried and sieved to 2 mm. Sieved soil (60 g) was packed into mini-lysimeters (8 cm height by 3 cm diameter) to a bulk density of 1.1 g cm^–3 (similar to field conditions; for method details, see Senbayram et al. 2015). Non-treated control was compared with three fertiliser treatments: dolomite, SMS and kieserite (single granule in each fertiliser treatment placed on top of the soil). In an additional treatment, fine-ground dolomite was applied to the soil surface. Three rainfall simulation events were applied by using a peristaltic pump, to simulate rain at 10 mm h^–1. Each event consisted of 30 mL rainfall within 3 h, and leachates were collected for Mg analysis. Three weeks after onset of treatments, soil exchangeable Mg at soil depth sections 0–2, 2–4, 4–6, and 6–8 cm were analysed by the CaCl₂ extraction method. Data from the experiment are presented in Figs 4 and 5, and clearly show no significant Mg leaching in either dolomite granule or dolomite fine treatment (known as non-soluble or slow release Mg fertiliser) compared with the non-fertilised control treatment, even after the third rainfall event (Fig. 4). Therefore, even fine-ground dolomite fertiliser does not cause detectable Mg leaching, even under extreme storm events. This is in line with the previous reports (Loganathan et al. 2005; Härdter et al. (2004). Soil analysis of exchangeable Mg showed that Mg concentration only in the top layer of soil (0–2 cm soil section) increased slightly in the dolomite granule treatment compared with the non-fertilised control soil. On the other hand, grinding dolomite caused greater release of Mg, and soil net exchangeable Mg content was ~4-fold higher in the dolomite fine treatment than the dolomite granule treatment. However, even in the dolomite fine treatment, total net exchangeable Mg in 0–8 cm soil section was 87% lower than with kieserite 3 weeks after the onset of treatments (Fig. 5). This result may suggest that dolomite or its ground form has very limited capacity to supply sufficient Mg under deficient conditions. When comparing soluble Mg fertilisers, kieserite and SMS differed greatly in leaching behaviour. Concentrations of Mg in the leachate of SMS-treated pots increased drastically (2.5-fold that of the kieserite treatment) after the first rainfall event (Fig. 4). However, after the second rainfall event, about twice as much Mg was leached in the kieserite treatment than the SMS treatment. Overall, 34% and 27% of applied Mg was lost via leaching in SMS and kieserite treatments.

Fig. 4.  Net magnesium leaching (% of applied fertiliser) after each rainfall simulation (30 mm each) in dolomite granule, dolomite fine (ground), synthetic magnesium sulfate (SMS) granule, and kieserite granule treatments. Fertilisers applied at a rate of 6 mg Mg lysimeter^–1. After each simulation, leachate was collected and analysed for Mg content. Data are means and standard errors four replicates.

Fig. 5.  Magnesium concentrations in various soil sections in soils with applied dolomite granule, dolomite fine (ground), synthetic magnesium sulfate (SMS) granule, kieserite granule and non-treated control at 3 weeks after fertiliser application. Fertilisers applied at a rate of 6 mg Mg lysimeter^–1. Three rainfall simulation events were applied, and after each simulation, leachate was collected and analysed for Mg content. Data are means and standard errors of four replications.

Compared with the above findings, Härdter et al. (2004) reported lower total Mg loss via leaching (16.7% and 22.3% of the applied Mg in kieserite and SMS treatment) from water-soluble Mg fertilisers. This was most likely due to differences in the size of the lysimeters. Soil columns were 20 cm high in the study of Härdter et al. (2004), whereas in the presently reported study, the lysimeters height was only 8 cm. However, both studies showed that Mg lost via leaching was significantly lower in soils with applied kieserite than SMS (21–33% lower). There is a clear indication that the more rapidly soluble SMS displaced larger amounts of cations from the exchange complex, causing greater losses of these nutrients than kieserite.

In this review, we have summarised current knowledge regarding the importance of Mg nutrition in plant growth and quality under changing climate and discussed the factors controlling Mg availability and leaching in soil. We can summarise three take-home messages.

First, the amount of Mg released from soil minerals is commonly not sufficient in sandy soils compared with the amounts needed to sustain high crop yield and quality. Especially in sandy soils, application of high rates of K or NH₄⁺ fertilisers often enhances the risk of Mg deficiency.

Second, Mg fertiliser supply in many agricultural systems is often inadequate mainly through lack of knowledge and/or economic reasons. Precise site- and crop-specific Mg fertiliser management practices need to be developed as recommended by IPNI 4R Nutrient Stewardship.

Third, unlike other cations, Mg is very mobile in soils because it is less bound to the soil charges. This results in a relatively high abundance of this element in the soil solution and thus higher risk of leaching. Mg-leaching properties of various water-soluble fertilisers may also vary depending on their chemical composition, granule size, and effect of soil pH and cation balance. Theoretically, the ideal Mg fertiliser would contain both high- and low-soluble Mg sources in agricultural systems where higher risk of leaching loss is expected.