Ecosystem as a Text: Semantic Analysis of the Global Vegetation Pattern

Let us consider a text, which is written by Russian language. At the first level of reception we know only a number of letters. Then the information per one letter I1=5bits. At the second level of perception we take into account the frequencies of letters then I2=4.35bits. At the next levels, when we take into account double, triple, etc. correlations, we get I3=3.5bits, I4=3bits, etc. Then the redundancy of information at each level, Ri=1-Ii/I1, is equal to R1=0, R2=0.13, R3=0.3, R4=0.4, etc. By defining the cost of information as a degree of non-redundancy, Ci=1/(1-Ri) we get C1=1 C2=1.15, C3=1.43, C4=1.67, etc. Let us consider now a description of the Global Vegetation Pattern (GVP). At the first level of description we have a number of biomes or vegetation types. In accordance with Bazilevich this number is equal to 30 then I1=4.9bits. At the second level we take into account the relative areas covered by biomes. Then I2=4.41bits and R2=0.1, C2=1.11. At the next level of description we consider the spatial correlations between different pairs of biomes (Bazilevich’s biomes map is used). We get the following results: I3=3.6bits, R3=0.265, C3=1.36. Semantic methods allow comparing two different texts, for instance, the GVP and the spatial distributions of temperature and precipitation.