The results confirm our first hypothesis that selective deployment of roots in high P patches than low P patches; thus root morphological plasticity plays an adaptive role in foraging heterogeneous P distribution by this species. The observed root morphological response is significantly modulated by patch strength and size; i.e. the greater the patch strength and the larger the patch size, the greater the deployment of root in high P than low P patches would be. The significantly higher root morphological traits on the high P side in the split-P treatment imply that P-deficiency signal from the low P side may stimulate the growth of the roots located in the high-P zone. This rooting characteristics increase the chance of encountering nutrient-rich patches, thereby enabling plants to efficiently forage in a heterogeneous soil profile [25]. The increase in root morphological traits over time with slight differences among levels of patch strength and patch size is mainly ontogenic difference but suggests that the plant adjusts it root system to meet the P demand as growth advances.
Spatial heterogeneity of nutrient in soils is reflected in the scale of the distribution. During the growth and development process, plant roots experience nutrient patches of different scales, and plant roots make corresponding morphological response [26]. Studies have shown that plant roots have a threshold for the “perception” of nutrient heterogeneity scale [27]. Under the condition of small heterogeneity scale, plants ignore the heterogeneity and regard it as homogeneous, and roots do not make plastic changes, or plasticity is very small. However, when the heterogeneity scale reached a certain threshold value, the root system cannot ignore the differences between heterogeneous environments, and a series of responses is triggered [27]. It has been shown that the difference in total root length in large, medium and small patches is not significant under the condition of high P, while the total root length in large patches can reach 3 times that of medium and small patches, and local root proliferation disappeared in small and medium patches under the condition of low phosphorus [1, 28–30]. Rhizomorphous clonal plants obtained nutrient resources by branching outwards, and the clonal plants were densely distributed in large patches with high nutrient content, while those in small patches were dispersed [31, 32]. This is in line with our findings that the root length, root surface area, root volume, and average root diameter in large patches were significantly higher than those in small patches.
In addition, the increased placement of roots in large patches with higher P concentration gradient could be related to the carbon cost for increased production of roots in smaller patches with low concentration gradient. Consequently, the plant has to make a “decision” on resource allocation to produce more roots in large patches to optimize nutrient capture, particularly for less mobile nutrients such as P. Interestingly all root morphological traits were low when the plant was initially grown in P-enriched patches than without initially encountering P (T9–T11 in Figs. 1–4). This indicates that N. reynaudiana is content with small amount of initial P availability, but slowly proliferated its root to forage patchily distributed P. As a whole, our results are consistent with previous studies that have demonstrated increased root deployment in localized nutrient-rich patches for a range of other species [5, 7, 8, 11, 17, 33, 34, 35].
The results also confirmed our second hypothesis that increased root P contents in high P- than low P- patch; suggesting that physiological plasticity plays an adaptive role in foraging patchily distributed nutrients in the soil by this species. Furthermore, the increased uptake of P in localized P-rich patches indicates that the species is highly efficient in resource acquisition. The fact that roots in patches with no P availability (0/30) had a certain amount of root P suggests internal redistribution of P to maintain P homeostasis. Similar results have been observed in low P tolerant Chinese fir genotypes [36]. While the translocation of P to the shoots and P content of leaf and stem were lower in plants initially grown in P-enriched patch followed by high P concentration gradient between patches, the P use efficiency was higher, especially in small patches. This indicates that this species has high P utilization efficiency when the availability of P is low. Utilization efficiency, defined as the amount of biomass per unit of nutrient present in the biomass, involves mechanisms such as remobilization of internal P, increased activity of enzymes that replace P in structural compounds or during metabolism [37, 38], or reduced consumption of P [39].
Total plant biomass is higher in heterogeneous than in homogeneous nutrient environment, but vary with patch strength and patch size, which confirms our third hypothesis. The total plant dry mass was significantly higher in large and medium patches in both high and low P concentration gradient between patches whereas plants initially grown in P-enriched patches produced more dry mass in large than small patch and homogeneous P distribution. This is in line with the general notion that foraging patchily distributed nutrients results in increased whole-plant productivity and growth rates compared with homogeneous environment [10, 11, 33]. Generally, N. reynaudiana produced less root than shoots (i.e. low root to shoot ratio) across all treatments; suggesting that the species has high P utilization efficiency, as also observed for other species [40]. Shoot biomass production was favored by availability of high concentration of P in the growing media. This is indeed expected as P is the main growth-limiting nutrients in the study area while it is essential for various plant metabolic processes. Thus, biomass production is more sensitive to heterogeneous than homogeneous P supply.