Introduction: Biodiversity in human-transformed landscapes
For centuries, global human-driven Land Use and Cover Change (LUCC) has spread the so-called ‘anthropogenic habitats’ in many regions, like the Mediterranean, thus determining biodiversity and ecosystem functioning in human-transformed landscapes (1 ). The impacts have been manifold, strongly differing from the mosaic landscapes of past and present organic mixed farming adapted to site-specific natural endowments, up to the land cover homogenization driven by industrial farming and animal feeding reliant on fossil fuels combined with spontaneous reforestation ensuing rural abandonment (2 ,3 ,4 ). Effects of LUCC on biodiversity are largely known (5 ,6 ), and they include a general decrease in species richness but also changes in species composition due to rarefaction of habitats specialists and expansion of generalists and cosmopolitan species, as well as invasions by alien species (7 ,8 ) through increasing propagule pressure and disturbance levels (9 ,10 ). All these impacts lead to biotic homogenization in most-human transformed regions (11 ).
Biodiversity components associated to human-transformed landscapes are clear examples of these changes. For example, population decline of most bird and butterfly species is believed to be linked to the loss and fragmentation of natural habitats, but also to the abandonment of traditional land-use management in rural areas that led to both agricultural intensification and forest encroachment (12 ,13 ). These LUCC effects added to other global change components, like climate change (14 ,15 ) combined with homogenization of cultural landscapes (16 ). Habitat specialists often are the most negatively affected, while generalists can even take profit of them (17 ).
As a result, human-transformed landscapes are the outcome of a shifting interplay between the spatial patterns of land-use types driven by the energy flows of human activity, and their associated ecological processes (18 ,19 ). A widely acknowledged consensus in conservation biology, framed in the intermediate disturbance hypothesis (20 ), points out that landscape heterogeneity is key to maintaining biodiversity peaks at intermediate disturbance levels through the interaction between the ecosystem patch diversity and the ecological requirements to activate the dispersal abilities of species that come from less disturbed patches or colonize the most disturbed ones (21 ,22 ). However, the outcome also depends on the intensity and spatial distribution of the anthropogenic socio-metabolic flows at play (23 ,24 ). This recurrent interaction between landscape patterns and socioecological pressures opens a research field on how the complexity of the energy flows driven by farming, livestock rearing, and forestry ‘imprint’ diverse spatial patterns of human land-uses that give rise to heterogeneous or homogeneous landscapes capable to house different biodiversity levels (25 ,26 ).
The fundamentals of this research agenda can be found in Morowitz’s theorem that says that a flow of energy through a system is a necessary and sufficient condition for generating an organized structure, although ephemeral in time (27 ). The structures of living systems that emerge and evolve towards a self-reproducing complexity allow to keep information organized and to transfer energy with higher efficiency away from thermodynamic equilibrium (28 ). Without doubt, this increase in internal complexity is achieved by exporting entropy to the environment, as any living organism is an energy dissipative structure with multiple metabolic cycles interrelated through a space-time heterogeneous base (29 ). This thermodynamic concept of organisms has close similarities to ecosystems’ sustainability, which is directly related to information-complexity and inversely to entropy. When a living system becomes more complex it is also metabolically more efficient because it increases the internal information instead of its energy intake, thereby also reducing external entropy (30 ).
Margalef principle indicates that ecosystems succession tends to a decrease in the photosynthetic Net Primary Production (NPP) growth rate. In other words, entropy combined with information increases diversity, not the production of uniformity (31 ). In the same vein, complex agroecosystems can store more energy and information at some points that reduce internal entropy, thanks to the exploitation of other spaces of lower complexity but larger NPP production rate within a joint encompassing landscape functional structure. Many traditional Mediterranean agroecosystems were the result of this type of balance between exploitation and conservation through the spatial localization of different gradients of human intervention, a wise intermediate disturbance pattern resulting in heterogeneous landscape mosaics (32 ).
The combination of an energy-flow pattern driven by complex information on how energy is redistributed across space is a good starting point for modelling the human-landscape relationships. According to Margalef (31 ), ‘the relationship between the external energy inputs and the dimensions that characterize the spatial patterns of its distribution ’ gives rise to the functional structure of landscape mosaics capable to host higher biodiversity than other homogeneous land covers. In order to test this hypothesis, Marull et al. (33 ,54 ) developed an Energy–Landscape Integrated analysis (ELIA ) of agroecosystems which accounts both the investment of energy stored within (E ) and the information held in the whole network of socio-metabolic energy flows (I ), to correlate their interplay (E·I ) with the functional structure of the resulting cultural landscape (L ). Then, we use here this energy-information-structure model (Fig. 1 ) to predict the locations of two important biodiversity components, butterflies and birds, in the human-transformed landscapes here accounted as an empirical example (31 ). The aim is to test Margalef’s hypothesis that the complex landscape mosaics of traditional organic agriculture were, and continue to be, good for biodiversity conservation. This means discussing if the energy reinvested and redistributed by farming-driven metabolic flows can lay the foundations to study the linkages that exist between social metabolism, landscape ecology and biodiversity, to better inform sustainable-oriented land use policies.
[Insert Figure 1 here]
Testing Margalef’f hypothesis requires specifying and accounting the ecological disturbance exerted by the information-driven external energy that farmers incorporate to the landscape. The Human Appropriation of Net Primary Production (HANPP) is a quantitative estimate of the potential annual biological productivity reduced by human activities (34 ,35 ). It provides a first approach to the interplay between the anthropogenic disturbances and the wildlife ability to withstand them. Intermediate HANPP values can maintain greater biodiversity in human-modified landscapes than higher ones that would largely decrease habitat differentiation and the provision of enough food chains free from human colonization (36 ). On this light, Marull et al. (33 ) developed an Intermediate Disturbance Complexity (IDC ) model to assess how different levels of anthropogenic disturbance at regional scale affect landscape functional structure. Results show a hump-shaped relationship between landscape complexity, free NPP available for non-domesticated species, and biodiversity levels (33 ,37 ). However, this strongly depends not only on the overall flux balance of photosynthetic biomass but on the intensity and distribution of socio-metabolic flows associated to either land-use mosaics or homogeneous land covers of agroecosystems (23 ,24 ,38 ).
The ELIA modelling goes a step forward from the previousIDC explorations of the links between intermediate levels of socio-metabolic human disturbance and the ecological functioning of cultural landscapes carried out at regional scale (37 ). It measures E as the quantity of energy remaining in the agroecosystem, and I as the complexity of the network of flows which allows farmers to reproduce the landscape fund components thanks to the information embedded in the system. Both indicators, E andI , bring to light how the energy-information interplay gives rise to human-transformed landscapes. They do so by relating them toL , the landscape ecology metrics that assess the ‘imprint’ of the energy flows driven by farmers. We surmise that the interplay betweenE and I jointly leads to complexity, understood as a balanced level of intermediate self-organisation (39 ). We also assume that the complexity of energy flows (E ·I ) is related to the landscape functional structure (L ). The cyclical nature of these matter-energy flows is relevant in order to grasp the emergent complexity and the greater information held within agroecosystems, since they imply an internal maximisation of some less-dissipative fractions of societal metabolism. The complexity and information carried out by these energy loops lay the foundations to better understand and develop more sustainable human-managed landscapes.