Modelling Aeolian Agro-Environmental Landscape Dynamics

Transcription

1 Innovation of Models across Agro-Environmental Scales: Modelling Aeolian Agro-Environmental Landscape Dynamics Andreas Baas Department of Geography King s College London thanks to:

2 Aeolian Agro-Environmental Landscapes Patrick Hesp Ted Zobeck Paul Gares

3 Agro-Environmental Modelling v Models are tools for understanding agro-environmental processes (e.g. plant production, plant-soil interactions, soil degradation and soil erosion) v Progressively used in land use planning and assessment v Need to reflect different spatio-temporal scales and complex interactions (e.g. socio-economic driving-forces, soil management practices and geoecological processes) v Challenges to: 1) cover processes across spatio-temporal scales 2) linking processes of different kind

4 Personal Modelling Background SAFE DECAL Baas (2002) Arens et al. (2001) Baas and Nield (2007)

5 Outline Goal: to review challenges and innovations - Historical context and types of models 1D, 2D, 3D - Challenges & opportunities scaling and vegetation - Recent innovations combination and versatility - Good modelling practices implications

6 Historical Context & Model Types

7 Dune Mobility Chepil-Siddoway-Armbrust (1962): Lancaster (1988): Thomas et al. (2005)

8 Dune Mobility Yizhaq-Ashkenazy-Tsoar (2009) Hugenholtz and Wolf (2005)

9 Agricultural Wind Erosion

10 Wind Erosion Equation, WEQ Woodruff and Siddoway (1965): E = ƒ (I, K, C, L, V) I = soil erodibility K = soil ridge roughness C = climate L = field length V = vegetation Nomograms: (graphical calculators)

11 Revised Wind Erosion Equation, RWEQ Fryrear et al. (1998):

12 Revised Wind Erosion Equation, RWEQ (85 parameters, coefficients, variables)

13 Wind Erosion Prediction System, WEPS Hagen (1991); Hagen-Wagner-Tatarko (1996)

14 Wind Erosion Prediction System, WEPS

15 Water Erosion Prediction Project, WEPP

16 Transect Models

17 Blow-outs Transect Models Drift sands Jungerius (1984) Jungerius, Castel Kootwijkerzand Exner and Namikas Castel (1988)

18 Transect Models Exner (1920), sediment continuity: transport gradients surface evolution Namikas and Sherman (1998)

19 Coupled Airflow Sand Transport Models

20 Airflow Modelling Van Boxel et al. (1999) Jackson and Hunt (1975) HILL/SAFE:

21 Arens et al. (2001) Simulating Aeolian Foredune Evolution, SAFE

22 Three dimensional

23 Three dimensional Two approaches: Reductionist physically-based coupled airflow-sand transport dynamics and resulting surface evolution Self-organising systems Non-linear dynamics, emergent behaviour, cellular automata (CA)

24 3D-Reductionist Narteau et al. (2009) Duran and Herrmann (2006) Hersen et al. (2004)

25 3D-Self-Organising Werner (1995) CA algorithm Wind direction erosion & transport

26 3D-Self-Organising Werner (1995) CA algorithm Wind direction deposition avalanching enforces angle of repose shadow zone 15

27 Werner Dune Model Bare sand dune landscape simulations Whole fields instead of single dunes

28 Werner Dune Model Bare sand simulations: barchan field evolution Baas and Nield (2005)

29 DECAL (Discrete Ecogeomorphic Aeolian Landscape Model) Wind direction deposition avalanching enforces angle of repose shadow zone 15

30 DECAL Wind direction dynamic conservative erosion growth deposition decline -100

31 DECAL Heather Viles

32 DECAL Nield and Baas (2008)

33 Applications to Agro-Environmental Systems 1D mobility regional scale & climate 1D WEPS agriculture, field unit soil erosion 2D SAFE 3D reductionist 3D CA coastal dunes physical principles (fundamental scales) organising principles, larger scales

34 Challenges & Innovations

35 Challenges & Opportunities Two challenges: Scaling Transcending and integrating across spatio-temporal scales Biological systems Incorporating and balancing flora and fauna (and humans)

36 Challenges & Opportunities Detail reductionist physically-based? Chaos

37 Challenges & Opportunities Detail reductionist physically-based? Single dune Dune field Duran et al. (2008) Nield and Baas (2008) Computation cost resolution & size

38 Challenges & Opportunities Flora and fauna? - What to measure? - How to quantify? - Feedback? M Salmon

39 Challenges & Opportunities Measurement & calibration problems Feeding more detailed process knowledge into models with limited measuring opportunities will not help to mitigate land degradation.

40 Challenges & Opportunities Answer? Keep It Simple

41 Innovations Gradient of model types - 1D 2D 3D - reductionist CA - discrete continuous - local global Innovations by moving along gradient or combining models

42 Map-Based WEPS Visser et al. (2005) Visser et al. (2005)

43 Good Modelling Practices

44 Good Modelling Practices validation parameter space exploration algorithm details & implementation scale & resolution uncertainty comparison fundamental & conceptual issues

45 Good Modelling Practices Nine Good Practices (Malamud and Baas, 2012): Model construction: 1) select appropriate model type/strategy 2) parsimony ( Occam s Razor ) 3) dimensional analysis 4) benchmark testing* Model running: 5) sensitivity analysis* 6) calibration 7) data exploration 8) uncertainty assessment 9) consider alternatives Each of these steps can lead to new insights!

46 Testing Algorithm: Directional Variability

47 Testing Algorithm: Directional Variability Probabilistic: horizontal/vertical as % of time to achieve an oblique x,y direction 60% 40% 2D transport paths: Displacement vectors with x and y components Rotation: Rotate domain and resample

48 Testing Algorithm: Directional Variability 2D path (chess) probabilistic time normal (L=5) normal (L=1)

49 Sensitivity Analysis / Calibration

50 Biggest Challenge

51 Final Thoughts

52 Final Thoughts Parameters and algorithm details need to be investigated thoroughly, These inquiries can lead to fundamental questions and insights, Many interesting science and application questions arise during model development, The journey is often more fruitful than the destination!

53 Final Thoughts