L-Band Radiometry Experiments wit ELBARA 1. ELBARA, te ETH L-band radiometer for soil moisture researc. Bare-Soil Experiment 3. Frozen-Soil Experiment 4. Clover-Grass Experiment 5. Forest Experiment 6. Summary 7. Outloo Overview Institutes: People: Hannes Flüler Massimo Guglielmetti Hannes Wydler Mie Scwan Cristian Mätzler 1
Tecnical Caracteristics: Dice type L-Band radiometer ( λ 1 cm, f 1.4 GHz) Two cannels (14-1418 MHz, 149-147 MHz) Dual polarized conical orn antenna (3.5 db gain) -3 db beamwidt of ± 6 Measuring cycle and operation: Full measuring cycle calibrations at T C =78K and T H =338K. T B at two cannels. Measurements are averaged during 1 seconds. A full cycle lasts for 6 s. Accuracy : Sensitivity: ±1K < 1K Automated rotation stage 3m 1.5m ELBARA, te ETH L-band radiometer for soil moisture researc
Wen: April June Were: Field-site close to Züric We found: none of te existing rougness models is capable to explain observed discrepancies between simultaneous measured ground trut and L-band radiometer data. Scope of rougness models: Fraunofer-criterion: Bragg-limit: σ < Λ> λ 3cosϑ λ sinϑ Te Air-To-Soil transition model considers te soil surface as a transition layer wit ticness ( = pea-to-pea rougness). Results from Bare-Soil Experiment In tis layer ε increases wit dept due to te factional increase of soil material. Te reflectivity results from a coerent, one-dimensional wave-propagation model. 3
Idea of te Air-To-Soil transition model: transition layer z*= z = z 1 z z z* z* 1 air macro pores soil material TDR averaging radiometer resolution order of β ( * ) soil ( ) soil ( * ) ( * ) in-situ probes ap p * soil β ε z = ν z ε z 1 ν z ε + air z* 1 β z* (Semi-empirical two-pase dielectric mixing model) Results from Bare-Soil Experiment z* ( ) ( ) soil * S z ( ) ν z = z d σ=.479 (for a Gaussian surface eigt distribution) * * 6z 6z * for z 3 + S z = for z < and z * * > 4
Fast Model for rougness correction wit low computational effort: Rroug ( Σ, ε, ϑ ) R ( Σ, ε, ϑ ) wit Σ=σ/λ=.479/λ Rspecular ( ε, ϑ ) multi-parameter regression: R(Σ,ε,ϑ) [-] 1..8.6 R ( Σ, ε, ϑ ) = exp ϑ =, pol. = p, s ε = 8 ε = Σ ( ε ϑ ) s,p a, H= / λ [-]..4.8.1.16..4. ϑ = 55, pol. = p ϑ = 55, pol. = s ε = 8 ε = 8 ε = ε =...1..3.4.5.6 Σ=σ / λ [-] a s,p s,p s,p ( ) ( ) εϑ, = 1 ϑ + ( ϑ ) ε s 4 8 1 s 4 8 p 4 8 1 p 4 8 ( ϑ) =.83663+ ϑ.813961+ ϑ.5938+ ϑ.85879 ( ϑ) =.714485+ ϑ.19949+ ϑ.43687+ ϑ.18565 ( ϑ) =.8417-ϑ.773684+ ϑ.381838-ϑ.71339 ( ϑ) =.785+ ϑ.955159-ϑ.1459+ ϑ. 1949 Results from Bare-Soil Experiment 5
r [-] r [-] Application of te Air-To-Soil transition model: a) Comparison between experimental soil reflectivity (blac) and calculated reflectivities derived from in-situ measurements wit = (gray) and =5mm (blue) respectively. b) Comparison between experimental soil reflectivity (blac) and reflectivities calculated considering a moisture-dependency opt (θ). c) Soil moisture θ in cm dept and opt (θ)..8.7.6.5.3.8.7.6.5.3 ( 1 ) T = T r B s radio r radio = 1-T B /T s r wit = mm r wit = 5mm r radio = 1-T B /T s r( opt (θ)) 16 165 17 175 18 day of year [d] opt (θ) [mm] 5 4 3 1.1.15..5.3.35 θ [m 3 m -3 ] Computed opt for total agreement between measured and calculated reflectivities plotted versus soil moisture θ in cm dept (dots). Linear fit opt (θ). Results from Bare-Soil Experiment 6
Wen: December Marc 3 Were: Field-site close to Züric Soil profile simulations: Using numerical soil-vegetationatmospere transfer model COUP. On day 4 soil started freezing to approximately 19 mm witin 7 d corresponding to an average frost front penetration velocity of 7.1 mm d -1. Comparison between soil reflectivities: Freezing results in reduced reflectivity. Frost dept [m] Snow dept [m] R radio and R COUP [%].3..1 -.1 -. -.3 8 6 4 19 mm no rougness correction (σ = ) Snow removed from te plot 7 d Snow removed from te plot frost penetration velocity = 7.1 mm d -1 4 6 Julian day wit rougness correction (σ = 5 mm) 1-1 - Air temperature [ C] Results from Frozen-Soil Experiment -1 1 3 4 5 6 Julian day 7
ε( z ) Estimation of soil-frost penetration velocity v: Measured reflectivities r during frost-penetration: leveled r radio [a.u.] 16 1 8 4 ε unfrozen = ε frozen = 4 τ =.9 d 9 1 11 1 Julian Day Hypotetical dielectric profiles: Temporal periodicity: τ =.9 d r(z) [%] 8 7 6 5 4 z v = = τ 7mm.9d = 5m md Calculated reflectivities r for penetrating frost: z = 7 mm 1 Spatial periodicity: z= 7 mm Results from Frozen-Soil Experiment 5 1 15 5 z [mm] 5 1 15 5 z front [mm] 8
Wen: June 4 August 4 Experimental Setup: Observation angles: ϑ= 45, 5, 55, 6 TDR and temperature measurements at, 4, 6, 1,, 3 and 5 cm dept. Were: Field-site close to Züric precipitation rate [mm/d] veg [cm] ρ veg [g/m ] 6 4 before ale: Field Conditions: 7 6 5 4 3 1 4 growt period 3 1 no vegetation longest dry period 1.7 cm/day 154 161 168 175 18 189 196 3 ale at day 19 86 g/m day day of year mowing at day 198 mowing at day 3 after ale: Results from Clover-Grass Experiment IR and precipitation measurements. 9
r s radio [-] r p radio [-].8.6...3..1. no vegetation L-Band Reflectivities : ( 1 ) T = T r B s radio growt period ϑ 1 = 45 α J = 5 α J = 55 α J = 6 154 161 168 175 18 189 196 3 day of year r radio ale at day 19 mowing at day 198 T = 1 T mowing at day 3 B s [-] r in-situ s [-] r in-situ p.8.6...3..1. In-Situ Reflectivities: no vegetation growt period ϑ 1 = 45 ϑ = 5 ϑ 3 = 55 ϑ 4 = 6 154 161 168 175 18 189 196 3 day of year r in-situ calculated from in-situ TDRmeasurements using a coerent, one-dimensional wave-propagation model. ale at day 19 mowing at day 198 mowing at day 3 Results from Clover-Grass Experiment 1
τ = γ H [-] τ v = γ v H [-] Opacities τ m = γ m H for te orizontal and vertical field-mode: 1..8.6.. 1..8.6.. no vegetation before ale: ϑ 1 = 45 ϑ = 5 ϑ 3 = 55 ϑ 4 = 6 clover-leafs: 1% random oriented grass-blades: 1% vertical oriented growt period 154 161 168 175 18 189 196 3 z day of year ail at day 19 mowing at day 198 mowing at day 3 z after ale: measurement based τ m : from non-scattering radiative transfer model + T B + in-situ measurements. T ( ) ( ) B = Tv 1 Γ + Tv 1 Γ r Γ ( 1 ) + T r Γ Modeled τ m : from dielectric mixing model applied to te assumed canopy structure. clover-leafs: 3% orizontal and 7% random oriented. grass-blades: 1% orizontal oriented s Results from Clover-Grass Experiment x x y y 11
Wen: October 4 October 5 Forest Canopy Transmittance Γ from upward measurement : From: T B + non-scattering radiative transfer model. Experimental Setup: { Γ or L-band Γ or X-band cumulative defoliation [%].3..1..3..1. 1 75 5 5 α = Alfa1 α = Alfa α = Alfa3 Were: Researc Center Jülic (Germany) α storm 8 9 3 31 3 33 day of year 4 Results from Forest Experiment (upward) 1
From Meteorological-Tower FZJ Experimental Setup: Results from Forest Experiment (downward) 13
From Meteorological-Tower FZJ T B - Angular Scan: Te observes T B K is in agreement wit expectations according to te non-scattering radiative transfer model: ( ) T T T =Γ T r r [K] T B or [K] T B or witout foil wit foil B B B v foil soil 7 65 6 55 5 45 4 35 8 75 7 65 6 55 T B = K T B = 1 K L-band witout foil during installation wit foil 5 4 44 46 48 5 5 54 56 58 6 angle [ ] X-band wit: r foil.9, r soil T = 8K, Γ v.5 Results from Forest Experiment (downward) 14
Experiment Bare-Soil Results Air-To-Soil transition model for describing te effect of small-scale rougness on te L-band emission. Summary Frozen-Soil Clover-Grass Possibility of monitoring a penetrating frost front into soil by L-band radiometry. Te L-band opacity of a vegetation canopy (clover-grass) is polarization dependent and correlated wit te canopy structure. Forest Te deciduous forest is semi-transparent at L- band. Te influence of te leaves on te transparency of a forest canopy is clearly smaller at L-band tan at X-band. 15
Next Experiments: Continuing T B - angular scan wit metal-foil until maximum foliation. Forest transmittances Γ for foliated and defoliated state. Outloo Continuing T B - angular scan witout metal-foil till fall. Observing dry and wet periods for answering te question of te measurability of soil-moisture troug te forest canopy under various conditions. Depending on te output of te previous experiments: Field-scale experiments wit artificially prepared forest soils (litter- or organiclayer, understory-vegetation, rougness ). Tower-experiment above an oter type of forest. Installing te instrument on an oil-platform. Any calibration / validation application for SMOS. 16