1. Mapping forest understory in European hemi-boreal stands using MODIS BRDF product Jan Pisek a , Miina Rautiainen b , Janne Heiskanen b , Matti Mõttus b July 26, 2011, IGARSS 2011 Vancouver a b

2. Why? the signal from the understory can be used, for example : (1) in forestry applications to separate forest site types (i.e. soil fertility classes) (2) to remove the influence of understory in estimating canopy biophysical variables (e.g. LAI, fPAR) from remotely sensed images (3) to develop and test canopy radiative transfer models Rautiainen et al., 2007, RSE

3. R n = P Tn x R T + P Gn x R G + Z Tn x R ZT + Z Gn x R ZG Is it possible?

4. R n = P Tn x R T + P Gn x R G + Z Tn x R ZT + Z Gn x R ZG R a = P Ta x R T + P Ga x R G + Z Ta x R ZT + Z Ga x R ZG R ZT = M R T ;R ZG = M R G Canisius and Chen, 2007, RSE; Pisek et al., 2010, IEEE-TGRS How?

5. R n = P Tn x R T + P Gn x R G + Z Tn x R ZT + Z Gn x R ZG R a = P Ta x R T + P Ga x R G + Z Ta x R ZT + Z Ga x R ZG R ZT = M R T ;R ZG = M R G (Peltoniemi et al., 2005, RSE) mixed understory

6. Bidirectional Reflectance Distribution Function (BRDF) The MODIS BRDF/Albedo Product (MCD43A1 ) at 500 m resolution every 8 days Photo property of Andres Kuusk

7. Hyyti ä l ä forest stands Herb-rich Birch Xeric Pine Mesic Spruce herbaceous species and graminoids lichens and heather m osses , abundant dwarf shrubs

8. Upscaling field measurements in Hyyti ä l ä area upscaling with forest site type database (produced by Metsähallitus)

9. Hyyti ä l ä results Field 2010 MODIS 2010 MODIS 2009

10. Hyyti ä l ä results – mean absolute errors DOY 100-200 (April-mid July) QA=0 QA=1 red band 0.007 0.006 NIR band 0.019 0.022 DOY 201-300 (mid July-October) QA=0 QA=1 red band 0.009 0.014 NIR band 0.045 0.058 DOY 100-300 (April-October) QA=0 QA=1 red band 0.008 0.011 NIR band 0.031 0.045

11. Finland understory NDVI - 2010

12. Hyyti ä l ä summary We retrieve seasonal dynamics of understory spectra at 500 m resolution with MODIS BRDF product Results directly validated with in-situ measured seasonal reflectance trajectories of boreal forest understory layers at Hyytiälä, Finland We illustrate the potential of MODIS multi-angle, multi-spectral observations to map understory properties, which was not possible with mono-angle data Always check and take into account documented quality and limitations of the satellite-derived product !

13. Want to know more? Pisek, J., Rautiainen, M., Heiskanen, J., Mottus, M. On the retrieval of seasonal dynamics of forest understory reflectance in a Northern European boreal forest from MODIS BRDF data. Remote Sensing of Environment (accepted with revisions). Rautiainen, M., Mõttus, M., Heiskanen, J., Akujärvi, A., Majasalmi, T. & Stenberg, P. 2011. Seasonal reflectance dynamics of common understory types in a Northern European boreal forest. Remote Sensing of Environment (in press).

14. RAMI stands (J ä rvselja, Estonia) RAMI Birch RAMI Pine RAMI Spruce Kuusk et al., 2009, RSE a mixture of several grass species various Sphagnum moss species and marsh tea bare soil, mosses

15. J ä rvselja results 36 (birch+alder) 51 (spruce) 13 (pine) 97 (pine) 3 (birch) 86 (birch, aspen, alder) 14 (spruce) Field MODIS 2007 Hazelnut LOPEX

16. RAMI birch stand (J ä rvselja, Estonia) abundant layer of hazelnut Field MODIS 2007 Hazelnut LOPEX

18. MODIS BRDF parameters (every 8 days at 500m resolution) R ( θ i ,θ v ,Φ,λ ) = f iso ( λ ) + f vol ( λ ) K vol ( θ i ,θ v ,Φ ) + f geo ( λ ) K geo ( θ i ,θ v ,Φ ) MODIS BRDF parameters from MCD43A1.005 product from Lucht et al. (2000, IEEE)