Jaques Baudry, INRA - National Institute for Agronomic Research, Jacques.email@example.com
Yousef Erfanifard, Shiraz University, firstname.lastname@example.org
Felix Herzog, Agroscope, email@example.com
Landscape ecologists rely heavily on mosaics to represent landscapes. In agricultural landscapes, the mosaic represents the different crops, or aggregations of land uses with no consideration of within category diversity. Three groups of methods can better represent landscape heterogeneity: 1) Pixel based mapping from remote sensing images can capture gradients such as differences in phenology, vegetation density, biomass etc. 2) interviews of farmers to know their practices (fertilization, pesticides etc. and 3) in situ spatial analysis of vegetation patterns.
The inclusion of vegetation structure and practices in models improves our understanding of biodiversity patterns and dynamics.
Landscape ecology is still a young science and a better characterization of landscapes improves our understanding of ecological patterns and processes as those related to biodiversity. The shift from matrix/ patch to mosaic representation has been a first step. The next should be representing landscapes as gradients (Cushman, Mc Garrigal, Lausch), the dominant paradigm in ecology. Maps of mosaic usually represent types of “land cover” such as woodland, grassland, wheat, etc. This hides both the heterogeneity within each patch and the heterogeneity among patches of the same type, due to e.g. management practices, physical environment. The objective is to bring together different approaches deciphering gradients at farm and landscape scales that improve our understanding of the dynamics of various forms of biodiversity (species of conservation interest or abundant species providing ecosystem services).
The presentations will address a landscape scale question that demonstrates how the consideration of heterogeneities at different scales (patch, cluster of patches, farms etc.) helps to explain biodiversity patterns. There is certainly a plurality of gradients that must be represented on different maps. With gradients, a single or few quantitative variables convey the information on the environment, a parsimonious way as compared to the combination of qualitative variables (patch types) whose biophysical differences are not assessed.
Remote sensing has long provided quantitative variables, as Leaf Area Index, meaningful to measure biomass. Novel high resolution radar images (SAR images, Sentinel1) or Lidar permit to go further in the analysis of landscape structure of perennial vegetation or annual crops, strongly improving our understanding of species distribution and population dynamics.
Independently of the mosaic of crops, crop management play a key role in controlling population dynamics. Within and between farms land use practices heterogeneities are largely overlooked by ecologists. The aim is to incorporate them as major components of landscapes.
Spatial statistics are the most reliable methods to analyze spatial structure of patches of plants at large- and small-spatial scales. Case studies in spatial structure of patches will convey a deeper knowledge about ecological processes at landscape scale. We will focus on: the quantification of spatial patterns of plants as the relationship between spatial patterns and processes in vegetation ecology.
Thence, the symposium will cover a diversity of approaches for discussing their complementarity and relative merits. We expect to bring together scientists with different backgrounds and fields of research, but all focused on landscapes and biodiversity. The objective is not only theoretical or methodological but also practical as land use practices and their effects on vegetation structure at patch and landscape scales at of overriding importance in the control of ecological processes.
What can participants expect to learn?
The participants will have a broad view of different ways to address, measure landscape heterogeneities at different scales. Beyond the technical aspects, they will discuss the conceptual and theoretical implications for mapping or building models.
As least, we will be able to produce a position paper summarizing the discussions and new directions for research. This will push forward the idea of representing landscapes as gradients already presented in some papers but not yet very much in practice.