Using a meta-analysis approach to understand complexity in soil biodiversity and phosphorous acquisition in plants

Current soil phosphorous (P) management is neither environmentally nor economically sustainable. Soil biodiversity has been offered as a solution to unsustainable land management and to promote ecosystem service provision. We know soil biology is instrumental in plant access to soil P, but specific effects of biological complexity, (used here to describe the number of links between different organisms), under different P levels on plant productivity are not well understood. We conducted a meta-analysis on relevant literature, which reported the response of terrestrial plants of economic and anthropogenic importance to P conditions, and controlled for biological treatments across different land-uses (arable, grassland and woodland). We hypothesised that: 1) in arable systems increased biological complexity will enhance plant productivity; 2) in perennial systems such as grassland and woodlands, increasing biological complexity will have no effect; 3) increasing the fertility of the system by addition of P fertiliser will reduce any benefits of biological complexity. We found that soil organisms are not always beneficial to plant shoot biomass, but that the effects of - and interaction among - bacteria, protozoa, nematodes, mycorrhizae, collembola and earthworms differ in their impact on plant biomass (positive or negative) dependent on the presence of other community members, P-level status and time. These findings bring into question existing frameworks that link below-ground biodiversity with above-ground plant productivity. We recommend further experimental work be conducted, which controls for land-use, P status, and soil biological composition and complexity. Such work should be followed by future systematic reviews, which could pragmatically inform more tailored biological management for plant P requirements, land-use and ecosystem service provision. To enable further meta-analyses of this type we recommend habitual inclusion of sufficient experimental detail and data, as a prerequisite for publication and a useful way to utilise increased online publication space.