Transcript Slide 1

Lichen height as a management
tool
in reindeer husbandry
Anna Olofsson1, Öje Danell1, Pär Forslund2 & Birgitta Åhman1
Reindeer Husbandry Unit1 & Ecology Department2, SLU, Uppsala, Sweden
Aim
To investigate how lichen height measurements should practically be designed in order
to get enough power to detect changes in
lichen cover.
Background
Lichen is a major food resource for reindeer and in
reindeer husbandry during the snow season and it is
important to avoid overuse of lichen resources. The lichen
is a slow-growing perennial resource and changes in lichen
biomass are difficult to monitor. However, knowledge of
ongoing change is essential in order to adapt the use of the
resources. A previous study showed that lichen height can
be used to estimate lichen biomass.
Material and methods
Lichen height was measured every 10 cm on 14 different 30 m transects
in lichen grounds with different grazing pressure and different amount
of lichen cover. Lichen height was also measured every 2.5 m in 6-12
transects of 100 m in 16 different grazing sites of good quality. In total
16 sites with 122 transects were measured in northern Sweden. Power
analyses, autocorrelation analyses and mixed linear models with
spatially correlated error terms were used to analyze the data.
Results
The autocorrelation disappeared within a few meters, although a cyclical pattern was found in some sites (Figure 1).
Within 10 cm autocorrelation varied between 0,35 – 0,76 with mean 0,55 and standard deviation (SD) 0,12. At a
distance of 20 cm the autocorrelation mean was 0,4 with SD 0,16. Forest stand age, forest density and site had
significant effects on lichen height in a model without interactions. Significant interactions between site and forest
density, and forest stand age and forest density was observable. Results from a preliminary power analysis are showed
in Table 1.
Table 1. Results from power analysis using
the mean standard deviation
Autocorrelation (ρ)
Mean
differenc
e
5
5
10
10
Distance (m)
Fig. 1. Autocorrelation pattern on the
small scale in three different sites.
The blue dots are autocorrelation values, the pink line is fitted
to the values with lowess, blue lines are upper and lower
confidence limits.
SD
19,98
19,98
19,98
19,98
Nominal
Power
0,9
0,95
0,9
0,95
Actual
Power
0,9
0,95
0,9
0,95
N Total
674
832
170
210
Conclusions
An effect of forest density was found, but varied
between sites. Therefore general adjustments for this
effect is difficult.
Number of measure points is a trade-off between the
needs of frequent detections of changes and trends
(early warnings) and the effort spent on measuring. The
longer time interval between measurements, the fewer
points has to be measured for detecting changes.
A solution might be to measure exactly the same spots
at every occasion, although since autocorrelation
disappears within short distance, the marking of the
measured spots has to be reliable.