The CAOS Project - Phase 2

DFG Research Group: From Catchments as Organised Systems to Models based on Functional Units (FOR 1598)

Within the 2nd phase of the CAOS research unit we work towards a holistic framework exploring how spatial organisation alongside with spatial heterogeneity controls terrestrial water and energy cycles in intermediate scale catchments. "Holistic" means for us to link the "how" to the "why" by drawing from generic understanding of landscape formation and biotic controls on processes and structures as well as to rely on exemplary experimental learning in a hypothesis and theory based manner. This also implies joint examination of soil, vegetation and atmosphere as coupled system rather than linear combination of different compartments. Our work is guided by the following reseacrh questions:

• How to define, characterize and link hydrological functioning and underlying controls across a hierarchy of spatial and temporal scales, including their interactions with the soil, vegetation and the atmosphere?
• Does spatial organization in catchments imply the existence of a hierarchy of functional units which act similar with respect to different hydrological functions and can we characterise their typical behavior in an exemplary manner?
• How to balance necessary complexity and falsifiability of catchment models to step beyond the input-output paradigm?
• What is the connection between thermodynamics, organizing principles, catchment structure and catchment functioning?

Figure 1. Catchment functioning reflecting context-dependent controls of different elementary functional units (EFUs) or lateral topological units (LTUs).

To mutually work towards this goal we composed 7 projects which closely cooperate within two overarching work packages:

• WP1: Linking hydrological similarity with landscape structure across scales
• WP2: Searching for appropriate catchment models and organizing principles

Within WP1 we further refine the existing stratified multi-method and multi-sensor setup to search for functional entities in the Attert experimental basin and, if they exist, to learn in an exemplary manner which structural features control their functional characteristics. This essentially includes identification of suitable metrics to discriminate functional and structural similarity from our data as well as identification of useful quantitative descriptors for the rather fuzzy term "hydrological function". Overall we aim to synthesize a protocol to decide "where to assess which data for what reasons" for characterizing hydrological functioning across a scale range of four orders of magnitude in a parsimonious manner.

Within WP2 we foster our distillery of parsimonious and nevertheless physically consistent model structures which rely on observable quantities and make use of symmetries in the landscape to simplify the governing model equations in a hypothesis based manner. To this end we compare concurring model structures (among those the newly developed CAOS model) and work towards a framework for an objective model inter-comparison with special emphasis on a) the added value of different data/information sources and b) on consistency of predictions with respect to distributed dynamics and integral flows.

Additionally, we aim in WP2 at linking the 'how' to the 'why' by synthesizing testable hypotheses that could explain whether spatial organisation has evolved in accordance with candidate organising principles. Ecology, fluvial geomorphology and thermodynamics offer a large set of such candidate organising principles for this issue. Based on our recent work we focus especially on thermodynamic limits and optimality principles like maximum entropy production, explore their value for uncalibrated hydrological predictions, and work out the necessary requirements on data and models for testing these principles. We put special emphasis on a possible experimental falsification of these candidate principles; also in close collaboration with the B2-Landscape Evolution Observatory in Tucson, Arizona.

To assure that CAOS contributes to a general advancement of hydrology, we will jointly synthesize our "scientific deliverables" to the hydrological community:

1. An improved understanding of how and whether landscape organization and hydrological functioning are connected, including quantitative descriptors for different hydrological functions;
2. A protocol recommending experimental methods to detect and characterize catchment structure and hydrological functioning across scales, ranked with respect to how much information they provide versus cost and labor;
3. A standardized framework for model inter-comparison with emphasis on the value of different information sources to judge consistency of model performance and concurring model structures;
4. A protocol to test how far organizing principles bear and apply, for instance for uncalibrated hydrological predictions.

For further information about the CAOS project please refer to our opinion paper Zehe et al. 2014 (HESS)

Further information about the achievements in the first phase of the project can be found here.