Sustainable intensification of agriculture is one promising way to meet the expected global increase in demand for food, fiber, fodder, and biofuel (Godfray and Garnett, 2014). However, edaphic stresses such as drought, soil nutrient availability, high acidity, and high salinity severely limit worldwide production. Managing nutrient deficiencies may be difficult, considering that the global efficiency of fertilizer application is frequently not more than 50% for nitrogen (N), less than 10% for phosphorus (P), and about 40% for potassium (K) (Fageria, 2012). Excessive fertilization may, in turn, promote groundwater pollution and gaseous N emissions. Hereby, the European Commission targets a 20% reduction in fertilizer quantities and a 50% reduction in nutrient losses by 2030 (European Commission, 2020).

Studies focusing on roots and on their role in nutrient acquisition are crucial to lay the basis of management strategies to increase crop production while improving resource use efficiency (Gregory et al., 2013). Root systems are strongly influenced by a wide range of abiotic factors such as gravity, soil compactness, soil water content, soil texture, aeration, nutrient availability, pH, and temperature (Yapa et al., 1988; Bengough et al., 2011; Kopke et al., 2015; Schneider et al., 2017; Hartmann et al., 2018; Correa et al., 2019; Hadir et al., 2021). Biotic factors (e.g., bacteria, fungi, nematodes, etc.) can also affect biogeochemical processes and affect the root morphology in the soil (Larsen et al., 2015). And vice versa, the root exudates stimulate microbial flora activity by fostering enzyme production. The microorganism decompose the soil organic matter, and consequently, the amounts of nutrients (N, P) increase, affecting the morphological traits in roots (Barrios-Masias et al., 2019)

Root systems can exhibit a high degree of plasticity in response to physical, chemical and biological changes in the environment (Lynch, 1995; Ostonen et al., 2007; Rich and Watt, 2013; Correa et al., 2019). For example, as reviewed by Correa et al. (2019), roots showed a retarded development as sign of apparent plasticity ¹ , including changes in architecture, as a response to severe stress (e.g. soil compaction). These architectural changes may in turn enhance the tolerance to variations in the environmental conditions (adaptive plasticity). Drew et al. (1973) showed that plants grown on nutrient-rich soil patches increased number and length of fine lateral roots, thus positively affecting the overall specific root length (SRL).

Gruber et al. (2013) grew Arabidopsis plants on agar at four deficiency levels for 12 nutrients and quantified seven root traits. Total root length increased by 48% under moderate N deficiency and decreased under most severe N deficiency. Furthermore, since the root biomass decreased comparatively less than the shoot, the root-to-shoot ratio gradually increased with decreasing N supply. In addition, N deficiency stimulated the growth of a more exploratory root system with long lateral roots. Foehse and Jungk (1983) reported that some N deficiency level stimulates root hair formation of spinach, tomato, and rape in pot experiments. Additionally, plants grown at low N displayed longer root hairs than plants grown at higher N concentrations. Moreover, when oilseed rape was grown in a split-pot system, root hairs did not form when all root system grown in media with poor N supply, whereas root hairs were formed when at least part of the roots (10%) was grown in N-rich media.

Crops cope with P deficiency by increasing root development in the P-rich zone (commonly in the topsoil) (Lynch and Brown, 2001; Rogers and Benfey, 2015), releasing carboxylates that capture iron and aluminium from the respective phosphates, thus rendering P more soluble (Hodge et al., 2009), as well as directing arbuscular mycorrhizal uptake pathways (Smith et al., 2018). Total root length generally decreases with P deficiency (Gruber et al., 2013; Haling et al., 2018), and the growth of primary and lateral roots is restrained when roots reach a low-P zone (Desnos, 2008). However, roots can also develop a shallower, horizontal, and highly branched root system (Lynch and Brown, 2001; Gruber et al., 2013; Müller et al., 2015). For example, beans develop more horizontal root angles under P-limited soil, resulting in a more extensive root area in the topsoil, where P was more concentrated than in the subsoil (Bonser et al., 1996). Another well-known mechanism to enhance the P acquisition in P-limited conditions is the increase in length and number of root hairs (Schmidt, 2001; Lambers et al., 2006).

In contrast to the numerous studies investigating root responses to N and P deficiencies, research on the effects of K deficiency in roots is scarcer. Notably, a study with Arabidopsis showed a decrease in root biomass (about 60%) and primary root length at the lowest supplied K concentration, while root-to-shoot ratios remained stable across different levels of K deficiency (Gruber et al., 2013).

The current understanding of root plasticity has been mostly derived from seedlings and pot experiments conducted in controlled environments such as greenhouses or phytochambers (López-Bucio et al., 2003; Rich and Watt, 2013). However, the root growth behaviors in those conditions are frequently different than those observed under field conditions due to several abiotic and biotic factors, which are more variable and differ significantly from those in the greenhouse (Rich and Watt, 2013; Watt et al., 2013; Heinze et al., 2016; Schittko et al., 2016; Rich et al., 2020). Plants growing in fields are usually grown in crop stands, thus interacting and competing with each other, changing their environment and that of their neighboring plants (Cahill et al., 2010; Faget et al., 2013; Weidlich et al., 2018). Thus, pot studies generally do not have the physical, chemical and microbial composition of field soils. This difference alters the growth rate and rooting depth of plants as compared to field studies (Eno and Popenoe, 1964; De Deyn et al., 2004; Passioura, 2006; Ruzicka et al., 2010; Ruzicka et al., 2012; Poorter et al., 2016; Howard et al., 2017). For instance, Mokany and Ash (2008) found a poor correlation of root biomass and root-to-shoot ratio in pot experiments vs. field conditions. Moreover, the root responses to any stress differ in pots compared to field, as shown in cassava, where the root weight and width were statistically similar under drought and irrigated conditions at field scale but different in the pot experiments (Kengkanna et al., 2019). Another limitation in pot studies is that the container shape affects root morphological characteristics. Roots of plants cultivated in smooth-sided containers can grow deformed or limit their growth because they cannot spread horizontally, as they would do in an open field, therefore, they expand vertically, wrapping up at the bottom of the pot (Amoroso et al., 2010; Oburger and Schmidt, 2016). Besides, the container influences the humidity, and ventilation of soil (Poorter et al., 2012). Consequently, transferring observations on root morphology or plasticity from pot experiments to real field conditions is usually impossible.

To overcome these limitations, we performed a systemic review and meta-analysis to analyze whether and how N, P, and K deficiencies impact root morphological traits of common arable crops under field conditions. We were particularly interested in root length, root biomass, root diameter, root hair formation and root/shoot performance indices such as root-to-shoot ratio, root length per unit of shoot biomass and specific root length.

We used a systematic review and meta-analysis approach to show the evidence of the effects of nutrient deficiencies on roots. The approach was as follows.

The spatial-temporal fluctuations and occurrences of nutrients in the soil are monitored by sensory mechanisms at root tips. This information triggers chemical signals which may shape root growth (Asim et al., 2020). The decrease in root length and root biomass upon N and P deficiency (see also Figures 1 - 3 ) seems to be a general property of root morphological plasticity. The low N and P availability negatively affects the above-ground part of the plant, including the leaf area and the photosynthetic capacity per unit of leaf area, consequently leading to a decrease in carbohydrates to be invested in root growth (Postma et al., 2014). Initially, a reduction in photosynthesis might be offset by an increase in the allocation of photosynthates to roots in order to maintain root growth. However, this resource relocation leads to a more pronounced shoot growth reduction, possibly limiting light capture and photosynthesis even more. Eventually, the smaller plants cannot sustain proper root and shoot growth, and absolute root length and biomass decrease.

Noteworthy, the above-mentioned general trend has exceptions. Some studies reported plants with longer roots in low nutrient conditions. In principle, the increase in root length could be a temporary effect in the early development stages (Peng et al., 2012; Xue et al., 2014). On the other hand, it could be that early investment in root growth under low nutrient conditions represents an advantageous strategy to cope with nutrient deprivation, e.g., as a tool to forage into the subsoil (Jia et al., 2022). Several reports indicate that the contribution of subsoil nutrients to overall uptake can be quite variable (Kautz et al., 2013) and also depend on other factors. Those include penetration resistance (Schneider et al., 2017), water distribution in the soil, as well as the availability of other nutrients, e.g., N abundancy when P is deficient (Bauke et al., 2017)

An apparent effect of N and P deficiencies on root morphology is the higher ratio between root length and shoot biomass. This may be explained by the enormous negative impact of N and P starvation on above-ground biomass, estimated at about 34% of shoot biomass decrease when N or P is deficient. Indeed, the root length also decreases due to the N and P deficiency, but not as much as the shoot biomass. Our review shows a decrease in root length of about 20% for N deficiency and 15% for P deficiency ( Figure 1 ), which is lower than the decrease in above-ground biomass.

Most of the root-to-shoot ratios ( Figure 2 ) increased under N deficiency. It is well known that the root-to-shoot ratio increases under N deficiency due to the concept of functional equilibrium. Competition for carbohydrates and nitrogenous compounds regulates root-to-shoot ratios. For example, when plants are changed from a non-N environment to an N environment with sufficient N supply, the shoot increases its growth in the short term, switching to a lower root-to-shoot ratio and delaying the root growth (Ågren and Ingestad, 2006). On the other hand, when the plant is transferred from a high N level to a zero N level, a non-equilibrium scenario appears; in the beginning, the ratio does not change much as long as free nitrate is available in the tissue, but when the internal nitrate content is depleted, the redistribution of organic-N determines the growth rate (Brouwer, 1983). In that scenario, root growth increases gradually more than shoot growth (Brouwer, 1983). In the end, shoot growth decreases when all the compounds are in N equilibrium. When the plants grow in a prolonged N-deficiency environment, the response to a renewed supply of N decreases (Brouwer, 1983).

Greenhouse (Horst et al., 1996; Shen et al., 2018) and lab studies (Rychter and Randall, 1994; Mollier and Pellerin, 1999; Ciereszko et al., 2011) have shown that the root-to-shoot ratio increases in low P conditions. However, our study could not confirm this finding, possibly due to the sample size (only three studies) which was too small to compare the effect between different conditions.

Specific root length was not affected by N or P deficiency consistently. For example, Ostonen et al. (2007) found a higher SRL in treatments with low nutrient levels. However, this finding was related only to the finest roots, and our review lacks the differentiation of root types. Poorter and Ryser (2015) have analyzed the response of specific leaf area (SLA) to light constraints and the specific root length (SRL) to nutrient availability constraints, as a similar response to constraints above and below ground crop parts, respectively. The changes in SRL were not as significant as SLA changes. However, by separating the root types by function (primary roots from lateral roots), the authors found that low nutrient levels positively affect the SRL of the lateral roots, which are supposedly most active in resource acquisition.

Due to a lack of data, our study can only conclude one consistent result with respect to the effect of K, which is the reduction in root length under K deficiency conditions. This observation can be explained, as in the case of N and P deficiency, with the lower availability of assimilates when K availability is reduced.

Interestingly, the root types monocot and dicot do not only share similar root morphology responses to N and P deficiency but also do so in similar magnitude (see Supplementary Figure S1 ), despite the differences in their root systems. However, some discrepancies exist in the relative change of root-to-shoot under P deficiency, which was be similar in dicot plants but greater in monocot plants compared with P-added soils. Under K deficiency, the data collected did not support a firm conclusion about the root morphology; however, the decrease in root length differed in magnitude between monocot (approx. 10%) and dicot (approx. 2%). It is similar to the study of Samal et al. (2010), who found a contrast in the magnitude of decrease among some crops tested under K deficiency. Therefore, despite the differences in the root architecture among crops and root types, it is highly likely that the fundamental regulators and sensing mechanisms are similar among monocot and dicot species.

To the best of our knowledge, this study considered all retrievable publications investigating root morphology in common crops at the field scale. Publications involved many soil types, weather conditions, management strategies, and genotypes. Furthermore, we showed findings contradictory to pot experiments and revealed the strengths of field-scale studies. Moreover, due to the meta-analysis of individual observations in each publication, we were able to quantify and statistically support the decrease in root length and biomass and the increase in root length per shoot biomass in low N and P environments. Our study had some limitations, though. None of the studies provided data on all the parameters we investigated. However, some studies had the data needed (such as root biomass, shoot biomass and root length) to calculate root-to-shoot data, root length per shoot biomass, and specific root length. We could calculate these ratios for a better comprehension of the deficiency response.

Nevertheless, the most critical limitation was the incompleteness of information about soil properties and nutrient concentration in the soils and crops in many studies. In this regard, our approach was to classify soil as “deficient” when the nutrient was not applied (0 kg ha^-1), which is not necessarily true depending on the soil nutrient content and the needs of a specific crop. Hence, the unfertilized treatment may or may not lead to nutrient deficiency.

Furthermore, our study did not address relevant interactions that may have an impact on the root morphology in the field, for instance, drought, soil temperature, and soil pH. They remain as open questions for further studies. Additionally, studies did not report about root-soil contact and interaction of roots with the rhizosphere microbiome and potential consequences for plant nutrient acquisition (Wendel et al., 2022) which remains a research gap.

Our study contributes to the knowledge about root adaptation to nutrient-deficient soils. We detected common mechanisms for how root morphology responds to N, P, and K deficiency, even though roots experience multiple interactions simultaneously in the field. Our main findings point out a decrease in root length and biomass but an increase in root length per shoot biomass and root-to-shoot ratio. These findings are particularly interesting for modelling of root growth and agroecosystem, which requires data about the changes in root traits under different nutrient conditions. Future work must now focus on elucidating interactions of nutrient-driven changes in root architectures with other environmental parameters, such as drought, temperature, the soil microbiome, or soil type. Particular focus could be lain on root nutrient plasticity at field scale, since its assessment with high temporal and spatial resolution is nowadays possible with the emerging non-invasive technologies for root phenotyping.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

GL, SS, JP contributed to the conception and design of the study. ASt, GL, SA, SS contributed to the search in scientific databases. GL organized the database, extracted the information, and made calculations and statistical analyses. GL, SS wrote the first draft of the manuscript. JP, WA, GS wrote sections of the manuscript. MA, FE, TG, MG, TK, SR, ASc, MW, PY contributed to improving and correcting the text and figures/tables. All authors contributed to the manuscript revision, read and approved the submitted version.