Index
Visualizing Data A-Z Single Raster Analysis Tools
Basic descriptive statistics with HISTO

1st Example

Let us start with a simple question:

given are a vegetation (raster) and a river (vectors) map.
What we want to find out: do certain plant associations cohere with proximity to rivers?

About the data:
the vegetation has been digitized from a analog copy. The original field data were superimposed on a 1:25000 map. The whole digital vegetation data exist in form of a Arc/Info® coverage. The import happened through Arc/Infos UNGENERATE format and IDRISIs ARCIDRIS module. As the latter produces vectors, a rasterizing process followed (resolution has been adjusted to 25 m). The river data originated from 1:50000 maps and were treated the same way as the vegetation information except the rasterizing.
Both data are georeferenced to the same coordinates.

  1. First we need to rasterize the river layer

    in IDRISI this requires a two-step procedure: (1) INITIAL to create a new empty image for the data to be filled in during rasterize with (2) LINERAS. If desirable IDRISI allows for copying spatial parameters (rows, cols, min x, min y, ...) from an existing image.

  2. Now have IDRISI calculate buffer zones

    the DISTANCE operator looks for non-zero cellvalues, takes them as a target and calculates the distances to these cells. The resulting image now holds euclidian distance values given in reference units. Use SPDIST where spherical distortions should be avoided. Keep in mind, that DISTANCE calculates for a 'flat' surface - results do not take into account slopes hence do not correspond to 'real' distances, especially in steep mountainous areas!
    The darker the color the closer we get to a water streamline. For a better orientation the vector layer rivers has been overlaid.

  3. Regarding our problem, which areas will be of interest?

    Consider the different data resolutions: vegetation on the basis of 1:25000, but rivers digitized from 1:50000 maps. The inspected region is part of the Alps, the rivers should better be delineated as brooks sometimes forming gorges with their own microclimate. For the sake of demonstration we will inspect the image for 100 resp. 300 m distance buffer zones around rivers.

    By applying RECLASS to the output image of DISTANCE (rivdist) we extract a 100 m buffer zone around the rivers: distance values between 0 and 100 are set to 1, all others to 0. The output image (here named riv100) contains only 1 and 0, 'yes' and 'no', therefor call it a boolean image

    We go on with OVERLAY, a very useful and often employed module when it comes to combine two images through addition, subtraction, multiplication, rationing, etc. The module takes the values of the two input images and proceeds with them according to the Overlay options writing the results to the output image cells:

    In that we multiply the vegetation with the boolean distance image, only those values (= vegetation classes) survive, that are multiplied with 1 - well, exactly these reside within our desired 100 m buffer zone. Black regions in the illustration below indicate areas outside the 100 m buffer.

  4. Calculating the area

    We need to stress the module AREA twice - (1) to know the overall area-size of each vegetation class (taurvege as input image for AREA) and (2) to get the areas in the buffer zone (veg100 as input). AREA offers three ways: create an image, where the cells of each vegetation class receive the summed area of that class, produce a values file or present the area for the classes as a table on the screen. In our case it seems appropriate to write out a values file. Further we must choose area measure units. I took kmē (number of cells, mē, acres, hectares, ... are among the choices).
    Repeat the steps starting with RECLASS now with a buffer distance of 300 m.

  5. Comparing with DATABASE WORKSHOP

    The DATABASE WORKSHOP should be subject of an own exercise, so only the principal line is sketched here.
    We ground the last step on three values files - that one for the areas of vegetation classes over the whole image, and the others showing the area sizes of the classes within the 100 resp. 300 m zone. Together with a values file of the class names a single database can be built up. The illustration below shows you, how additional database fields can be calculated from existing ones by standard SQL-statements (here the percentage of area within buffer zones related to overall area).

    Very high percentages mean, that most of the area of that vegetation class resides within a 100 resp. 300 m distance buffer zone. Notice that water does not occupy 100 percent! This fact may well result from the data heterogeneity. Obviously the vegetation map contains a small lake, etc. that is not part of the rivers dataset.

    Interpretation and ingenuity of the results in detail may be ceded to experts, but without knowing anything about the vegetation classes, we could make careful assumptions about which associations could be influenced by proximity to waters.

    IDVegetation Class NameOverall areaArea within 100 m zone [kmē]Area within 300 m zone [kmē]Percentage 100 m zonePercentage 300 m zone
    1Androsacion alpinae56.986873.69312513.66756.48065923.98359
    2Salicion herbaceae6.8731251.0831253.16515.7588446.04892
    3Caricion curvulae 15.144371.7418755.87062511.501838.76439
    4Seslerio-Semperviretum0.446250000
    5Aveno-Nardetum14.874381.9043756.30812512.8030642.40934
    6Agristio-Trifolio-Deschampsietum cespitosum5.3706251.908753.7535.5405669.82427
    7Agrostio-Trifolio-Deschampsion2.35250.8531251.952536.2646182.99681
    8Polygono-Trisetion0.510.3243750.5163.60294100
    9Dactylo-Poion0.0231250.0056250.02312524.32433100
    10Brachypodio-Koelerietum0.02187500.0075034.28571
    11Loiseleurietum3.270.496251.5687515.1758447.97401
    12Rhododendretum ferruginei5.6206250.942.71812516.7241248.35984
    13Erico-Rhododendretum hirsuti0.003750000
    14Junipero-Callunetum0.7218750.0550.257.61904834.63203
    15Pinetum mugli1.4068750.1143750.391258.12972127.80986
    16Alnetum viridis5.6168751.176253.2812520.9413658.41771
    17Cembretum0.4493750.0268750.2143755.98052847.70515
    18Larici-Cembretum1.6593750.10750.5856256.47834335.2919
    19Vaccinio-Rhododendro-Laricetum0.1981250.0206250.062510.410131.54574
    20Larici-Piceetum4.8843751.441253.677529.5073675.29111
    21Piceetum subalpinum0.0062500.006250100
    22Luzulo-Piceetum2.35750.8293751.84687535.1802778.3404
    23Alnetum incanae0.1606250.10.16062562.25681100
    24Salicetum eleagni0.0531250.0250.05312547.05882100
    25Adenostyletalia1.0481250.13750.60312513.1186657.54323
    26Dryopteridetum0.3231250.0718750.25437522.2437178.7234
    27Caricion fuscae0.4756250.281250.44437559.1327293.4297
    28very moist soils and fountain areas0.3018750.1943750.30187564.38924100
    29Glaciers11.60750.1368750.7231251.1791946.229808
    30Waters0.82750.62750.7862575.8308295.01511
    31Other regions (scree flora or cultivated areas)0.4043750.0543750.0812513.4466820.09274


Index
Visualizing Data A-Z Single Raster Analysis Tools
Basic descriptive statistics with HISTO
last modified: | Comments to Eric J. LORUP idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis idrisi resource center idrisi gis idrisi gis