To the list of courses || RASA2018 || To the theme || Estonian

**Data**: Sadametsapunkti.xls ### Instructions

## Exercise 3966. Points 3, theme: Spatial Autocorrelation |
Open exercise |

The attached file contains data of 100 forest sites observed in Saare County in 2002: coordinates of observation sites, coverage of tree species (from *kuusk* to *muu*), stand total coverage (*katvus*), land cover type (*maakate*) and reflectance values measured by the Landsat ETM+ (TM1 ... TM8).

- The stand composition of closer stands is more similar. Up to which distance the simlarity holds?
- How strong and how significant is the Mantel correlation between distance and stand composition within radius 500 m and within 1000 m?
- Up to which distance is the correlation statistically significant at p < 0.05?

This exercise can be solved using the online calculator.

- Find Mantel test and read how to mount the data. The data must be distances (or similarities) in pairs in one aspect and in another aspect. The correlation between these distances is calculated. The distances needed for Mantel test can be calculated in similarity module of the online calculator. Notice that unique ID numbers must be in the first column of the cases prepared for distance calculation.
- Calculate geographical distances and distances in stand composition from data in Sadametsapunkti.xls. Save the results.
- Open the results in Excel and copy the geographical and stand distances to be in adjacent columns and the distances within the same pairs to be side by side in t in the same row.
- Copy the data block to the input of the Mantel test of the online calculator.
- Press Calculate.
- The graphs should look similar to the example in the figure on right.
- See once more what was asked and try to answer the questions.

Log in to send your results and to see the expected answer and responses from other students.