Modelling the handshaking between atmosphere and subsurface at the root-zone interface

Transcription

1 Mitglied der Helmholtz-Gemeinschaft Modelling the handshaking between atmosphere and subsurface at the root-zone interface Jan Vanderborght

2 Overview Root water uptake in simulation models R-SWMS model: coupled 3-D soil-root system model Effect of root architecture, root hydraulic properties Water uptake from saline soils Hormonal versus hydraulic regulation of transpiration Upscaling

3 Background Trenberth et al., (2007)

4 Decay of Transpiration During Dry Spells DOY ET = ET 0 exp t t 0 λ Teuling et al. 2006, GRL

5 Effect of Root Water Uptake Model on Surface Water Flux Temperate humid climate Semi-arid climate Prediction of the decrease of evaportranspiration with time by different LSMs effect of uptake by deep roots? Teuling et al. 2006, GRL

6 Representation of Land Surface-Atmosphere Interactions Vertical scale is the scale of the canopy and soil profile: O 10 1 m Vertical resolution: O m

7 Representation of Root Water Uptake in Land Surface-Atmosphere Interactions Root length density Lr

8 Why is there evergreen forest in regions where the dry season rainfall is below 1.5 mm d -1? Nepstad et al., Nature, 1994

9 Nepstad et al., Nature, 1994

10 Representation of Root Water Uptake in Land Surface-Atmosphere Interactions Root length density Lr Volumetric water content Water sink term

11 Representation of Root Water Uptake in Land Surface-Atmosphere Interactions Root length density Reduction of uptake due to lack of water Lr Root water uptake compensation Volumetric water content Water sink term

12 Root Water Uptake Compensation Markewitch et al., New Phytologist, 2010

13 Hydraulic Lift Katul and Siqueira, New Pythologist, 2010.

14 Hydraulic Redistribution/Lift Neumann and Cardon, New Phytologist, 2012

15 Hydraulic Redistribution/lift Ecological functions of hydraulic lift increasing dry-season transpiration Providing water to shallow rooted plants Moistening upper soil layer to keep up nutrient uptake and microbiological activity prolonging life span of fine roots and maintaining root soil contact in dry soils moving precipitation down into deeper soil layers

16 Partial Root Zone Drying (PRD) Effect of heterogeneous water distributions in horticulture, orchards.

17 RWU in Vadose Zone Hydrological Models Soil-based approach: sink-term, S(z,t), with simplified assumptions regarding RWU t z s K( s ) Sz, t z Several properties or characteristics of root systems are not represented or considered

18 RWU in Vadose Zone Hydrological Models S is proportional to the root length density S x T p R RLD RLD x x dx T p b(x) From Hydrus manuals Simunek et al.

19 RWU in Vadose Zone Hydrological Models Stress response functions are used to account for water stress on RWU: S x T b( x), p T p Feddes and Raats, 2004 van Genuchten, 1987

20 Ro Coupled soil and root water flow Solute transport and uptake (large scale) Data comparison Root growth Solute transport and uptake (small scale) Solute transport inside roots/signaling

21 Water Flow in a Root System (Doussan et al. 2006) Hypotheses: osmotic gradient is negligible/ no capacity/ S-S conditions for each root node: x( z) J x K x Ax dl J r K r S r seg ( z) ( z) System of equations with x as unknown BC: J x (t) or x (t) at the collar s for each node s x s s s collar J r x J x J r x J x J r x J x J x s x J r

22 Coupled Water Flow Model R-SWMS (Javaux et al. 2008) SOIL PLANT ROOT geometry: grid with hexahedra subdivided in tetrahedra geometry: tree-like structure with connected nodes Root collar numerics: Richards equation with sink term S given by the soil-root fluxes. Based on SWMS_3D (Simunek, Huang, and van Genuchten, 1995) S V J s r numerics: system of linear equations with boundary conditions given by the soilroot interface water potential s and the plant collar /flux time series

23 How do Stomata Sense Soil Water Potentials? gs (stomatal conductance) leaf hormone concentration Hormonal signal Hydraulic signal plant hormones root soil

24 Upper Boundary Condition Stress function is the ratio of the transpiration rate of the plant compared to the transpiration rate if there would be sufficient water available.

25 Modeling Root Water Uptake with R-SWMS Köbernick et al., Frontiers in Plant Sciences (in press)

26 Effect of Root Phenotypes on Root Water Uptake Leitner et al., 2014, Field Crops Research

27 Effect of Root Phenotypes on Root Water Uptake

28 Effect of Salinity (Osmotic Head) on Root Water Uptake Water flow across root: J r K S (( zz) ) ( z( )) z) r r r ( ( z) r ss O x x J x Water flow through xylem + o Y s J r x collar J x K x A x x ( z) dl seg + o J x + o s : water potential (surface soil, xylem, osmotic) s J r x J r K K S r A r x : intrinsic radial conductivity : intrinsic axial conductivity : root -soil interface area x : xylem cross sectional area + o s J r J x x J x x Agrosphere (IBG-3)

29 Effect of Salinity (Osmotic Head) on Root Water Uptake/ Transpiration Δ ( s leaf ) hpa Δ ( o + s leaf ) hpa Schröder et al., Plant and Soil

30 Effect of Salinity (Osmotic Head) Combination with Matric Potential Stress?, o, o?, min, o or or o o o o o = T act /T pot

31 Simulations with R-SWMS: Single Root Soil-root interface low Tp low salinity Bulk soil high salinity high Tp Schröder et al. Plant and Soil (2013)

32 Simulations with R-SWMS: Single Root

33 Hormonal Signaling

34 Hormonal Signaling and Partial Root Zone Drying Huber et al., Plant and Soil (2014)

35 Hormonal Signaling and Alternated Root Zone Drying

36 Hormonal Signaling and Alternated Root Zone Drying

37 Using R-SWMS Simulations for Upscaling (Couvreur et al., HESS, 2012) A simple approach to consider root hydraulic properties and root architecture in larger scale simulation models T S a root K rs R x T SSF( x) SSF( x) K ( x) a root collar SSF( x) ( x) dx s comp s root K rs : root system s conductance SSF: standardized sink fraction (uptake for uniform water potential) collar : water potential at the collar root : water potential felt by the root system For a given collar, T a does not depend on T p S(x) is non-local: depends also on at other locations S(x) can be negative roots exude water hydraulic lift.

38 Relation between SSF and Hydraulic Properties of the Root System SSF is different from RLD SSF depends on hydraulic root parameters Couvreur et al., 2012, HESS

39 Simulation of RWU Compensations RWU compensation does not require a complex parameterisation. Occurence of RWU compensation does not depend on mean root zone water potential not related to local water stress. Also exudation is predicted (hydraulic lift) Couvreur et al., 2012

40 Upscaling from 3-D to 1-D SSF up x = 1 Ω up Ω up SSF x dx ψ root,up x = Ω up SSF x ψ s x dx Ω up SSF x dx S up x = T act Ω up Ω up SSF x dx + K comp Ω up Ω up SSF x ψ s x ψ root dx S up x = T act SSF up x + K comp SSF up x ψ root,up x ψ root Problem: how to compute ψ root,up x and ψ root when information about small scale variation of ψ s x is not available?

41 Upscaling Assumptions 1 SSF x ψ Ω s x dx 1 SSF x dx 1 up Ω up Ω up Ω up Ω up Ω up ψ s x dx ψ root,up x 1 Ω up Ω up ψ s x dx = ψ s,up x ψ root 1 Ω R Ω R SSF up x ψ s,up x dx t z s, up K( s, up ) Sup z, t z S up z = T act SSF up z + K comp SSF up z ψ s,up z ψ root

42 1-D 3-D Upscaling from 3-D to 1-D: Wheat ψ root,up = Ω SSF x ψ root x dx K comp SSF up x ψ root,up x ψ root Ω SSF x dx ψ root,up ψ s,up = 1 Ω Ω ψ s x dx K comp SSF up x ψ s,up x ψ root Couvreur et al., 2014, HESS

43 Upscaling from 3-D to 1-D: Maize with Row-Interrow Variability?

44 Hormonal vs. Hydraulic Signaling and Plant-Scale Behavior Two plant types: Isohydric plants: stomatal conductance depends on leaf water potential and on hormone concentrations in the leaves Anisohydric plants: stomatal conductance depends only on hormone concentrations. Different scenarios: Root fraction in wet/dry soil Transpiration rate

45 Tact/Tpot Hormonal vs. Hydraulic Signaling and Plant-Scale Behavior Isohydric Anisohydric Huber et al., 2015, Plant and Soil

46 Some Final Statements Models should be as simple as possible but not simpler. Albert Einstein Lack of data must not be a motivation to simplify models it should rather be a motivation to improve our experimental methods.

47 THANK YOU! Gaochao Cai Natalie Schröder Katrin Huber Elien Kerkhofs Mathieu Javaux Andrea Schnepf Helena Jorda Andreas Pohlmeier Asta Kunkel Magdalena Landl Felicien Meunier Betiglu Abesha Valentin Couvreur