This page provides descriptions of the architectural, spectral and illumination related properties of a 124 year old Pinus sylvestris stand located at 58° 18′ 47.13″ N 27° 17′ 48.23″ E, The stand was inventoried in the summer 2007 by Andres Kuusk, Joel Kuusk, Mait Lang, Tõnu Lükk, Matti Mõttus, Tiit Nilson, Miina Rautiainen, and Alo Eenmäe of the Tartu Observatory, in Tõravere, Estonia as well as the Estonian University of Life Sciences, Tartu, Estonia.
Potential RAMI participants thus are to treat the information presented on this page as actual 'inventory data', that is, they should identify/extract those parameters and characteristics that are required as input to their canopy reflectance models.
In some cases this may mean that simplifications have to be made to the available information, or, that parts of the available information cannot be - or have to be modified before being - exploited with a given radiative transfer model. Whatever the case may be, all potential RAMI participants should mimic the standard practices that they use when matching actual field measurements to the required set(s) of input parameters for their model(s). If this means that you need more information than provided, please do not hesitate in contacting us. Last but not least, for those 3D models capable of maintaining architectural fidelity down to the individual shoot and branch level a series of ASCII (text) files containing the Cartesian coordinates of various geometric primitives (triangles, spheres and cylinders) and their transformations will be given.
The Järvselja Scots Pine forest inventory was carried out over a 100×100 m² area placing the origin of the coordinate system at its south-western end. In order to include also the tree crowns of the inventoried tree locations within the RAMI Scots Pine Summer stand representation it was necessary to expand the scene area slightly beyond one hectare. Maintaining the origin of the tree location coordinate system thus resulted in some negative x,y values in the table below. Overall architectural characteristics of the scene are thus as follows:
Scene dimensions:( X × Y × Z)
105.932 × 106.118 × 18.560 [m × m × m]
(Xmin, Ymin, Zmin)
−2.773, −3.623, 0.0 [m, m, m]
(Xmax, Ymax, Zmax)
103.159, 102.495, 18.560 [m, m, m]
Number of trees in scene
1120 (1114 pine, 6 birch)
Leaf Area Index of scene*
2.3020 (2.2881 pine, 0.0139 birch)
Fractional scene coverage**
*The LAI of the pine trees is computed using half the total area of the needles in a shoot.
**The fractional cover is defined as 1 - direct transmission at zero solar zenith angle.
The table below provides the structural characteristics of the Scots Pine shoots (left) and the birch leaves (right).
Individual shoots of the Scots Pine trees are generated in using most of the properties presented in Table 1 of Smolander et al., 2003 (RSE).
The shape of individual birch tree leaves is approximated from photographs as depicted in the right hand picture.
RT models capable of representing the architecture of individual foliage elements with a series of geometric primitives (triangle, sphere, cylinder) may want to use the information provided in the ASCII (text) files accessible from the last row in each table below.
This total needle area value arises if the needles are represented as elongated spheres (as is the case in the ASCII file accessible via
the link in the last row of the above left-side table). If individual needles are represented as cylinders (with discs as endcaps) then the
total needle area of the shoot is 159.03 cm². The number of shoots per pine tree should be adjusted accordingly.
The Järvselja Scots Pine forest is generated on the basis of 11 individual tree representations. Ten of these pertain to the Scots pine (Pinus Sylvestris) species and one refers to Birch (Betula Pendula). The table below provides an overview of some structural characteristics of these 11 tree representations. For those RT models capable of representing the 3D architecture of a given tree through a series of geometric primitives the last lines of this table contain links to data files with detailed specifications of the foliage and wood structural properties of the Järvselja Scots Pine forest (summer) trees.
=: for shoots the zenith angle of the foliage normal is defined as the angle between the vertical and the normal of the inner/main twig of the shoot (for a shoot axis aligned along the z-axis the normal was arbitrarily chosen to lie along the y-axis). Rather than spanning the full range of possible zenith angles (i.e., from 0 to 180 degree) as could be expected for non-flat asymmetric objects, it was chosen to follow the convention of foliage normals pointing only into the upper hemisphere. This is because RAMI participants, that make use of this foliage normal distribution information, will in all likelihood have models where scatterers are represented as flat (disc or equilateral triangle shaped) objects. However, should your model require a description of the foliage normal zenith angle distribution up to 180 degrees then please do not hesitate in contacting us and we will provide this information to you. For both the zenith and azimuth angle distributions the 'graph' link shows an image of the normalised foliage normal distribution versus zenith (or azimuth) angle of the foliage normal. The 'data' files for the zenith and azimuth angle distribution have three columns indicating
the upper value of the zenith (or azimuth) angle in a given bin,
the fraction of foliage area having a normal in this zenith (or azimuth) angle range,
and the fraction of wood area having a normal that falls in this zenith (or azimuth) angle range. Bin angle widths were chosen to be 5 degrees and 10 degrees for zenith and azimuth angles, respectively.
x: the crown radius of actual trees is varying with azimuth angle. This can be seen in the various pictures showing a perspective-free nadir view of a given tree located at x=0,y=0 (concentric circles indicate the distance from the origin in steps of 0.25m). The mean and maximum values were computed from the triangle objects making up the 3D trees depicted in the picture in the the fourth-last row of each table column.
*: the graphs show the maximum radial distance of foliage elements in a given height interval plotted against the upper height level of that height interval. The data files have five columns: lower-height-of-bin-in-units-of-meters upper-height-of-bin-in-units-of-meters minimum_radial-distance_of_foliage-in-units-of-m maximum_radial-distance_of_foliage-in-units-of-m. mean_radial-distance_of_foliage-in-units-of-m
#: this value corresponds to the one-sided leaf area for flat leaves. For Scots pine trees it corresponds to the sum of the (maximum) silhouettes of all the individual needles in the tree (i.e., half the total needle area per tree).
°: the data files have 3 columns: lower-height-of-bin-in-units-of-meters upper-height-of-bin-in-units-of-meters area-of-wood-or-foliage-in-units-of-m2.
+: his is the nominal value derived from the inventory data for a tree of this height. The actual value of the tree representations provided in the ASCII files at the bottom of this table might be slightly different.
The Järvselja Scots Pine forest is composed of 1120 individual trees. The following table indicates how these trees are distributed among the above tree classes and specifies their respective x,y locations of the tree centers of each tree class in the forest stand. The last row of this table contains an ASCII file with tree rotation and translation information for those RT models capable of ingesting the detailed 3D architecture of the tree models specified in the previous section.
x: these files contain pseudo code to rotate individual trees around their z axis and translate them from the origin to the x,y locations specified in the data files of the previous row of this table. Positive rotation angles in these files indicate that when looking down from the positive Z axis towards the origin of the coordinate system a counterclockwise rotation will result in moving the positive x axis towards the positive y axis. The angle of rotation is in the 7th column of these data files (starting the count from 1).
RAMI participants with 3D RT models capable of representing objects using geometric primitives can download a single compressed ZIP archive with all the tree architectural ASCII information that is listed in the above tables by clicking HERE. Note: The size of the compressed archive is about 9.3 megabytes. It contains 46 ASCII files and can be unzipped using 'WINZIP' on windows or 'unzip' on linux/unix operating systems. Beware that the inflated archive will take up 65.9 Megabytes of storage.
Spectral canopy characteristics
All of the foliage wood and background components in the Järvselja pinestand (summer) scene feature LAMBERTIAN scattering properties. The tables below contains the magnitudes of the reflectance and transmission characteristics of the various canopy components for nineteen different spectral bands. The experimental identifier for the Järvselja Pine Stand (Summer) scene is given by HET07_JPS_SUM_B**_36 where B** relates to the spectral bands (B01, B02, …, B19). An ASCII (text) file that resumes all of this information can be found here.
The illumination conditions for the Järvselja Scots Pine forest stand relate to the 10th June 2005 at GMT 9:42-9:44. More specifically illumination contains both a direct and an isotropic diffuse component. The direct solar illumination is characterised by a solar zenith angle (SZA) of 36.6 degree and a solar azimuth angle of 299.06 degree. The table below indicates the ratio of isotropic diffuse to total incident radiation for the nineteen different spectral bands: