Table 1 Fitted binary elevation vectors onto non-metric multidimensional scaling (NMDS) ordination of species presence/absence matrix using the Bray-Curtis similarity index.

Elevation class	NMDS1	NMDS2	R2	P value
500 m	−0.67523	−0.73761	0.1782	0.072
550 m	−0.67523	−0.73761	0.1782	0.072
600 m	−0.90268	−0.43032	0.3813	0.002
650 m	−0.90268	−0.43032	0.3813	0.002
700 m	−0.90268	−0.43032	0.3813	0.002
750 m	−0.90268	−0.43032	0.3813	0.002
800 m	−0.90268	−0.43032	0.3813	0.002
850 m	−0.90245	−0.4308	0.4424	0.001
900 m	−0.97197	−0.23509	0.3998	0.001
950 m	−0.97197	−0.23509	0.3998	0.001
1000 m	−0.93194	−0.36261	0.3856	0.001
1050 m	−0.94818	−0.31773	0.3886	0.002
1100 m	−0.84171	−0.53993	0.4012	0.003
1150 m	−0.84171	−0.53993	0.4012	0.003
1200 m	−0.80926	−0.58745	0.4246	0.001
1250 m	−0.70451	−0.7097	0.4480	0.001
1300 m	−0.72360	−0.69022	0.4177	0.001
1350 m	−0.64759	−0.76199	0.3666	0.002
1400 m	−0.64759	−0.76199	0.3666	0.002
1450 m	−0.67400	−0.73873	0.2786	0.007
1500 m	−0.67400	−0.73873	0.2786	0.007
1550 m	−0.67400	−0.73873	0.2786	0.007
1600 m	−0.67400	−0.73873	0.2786	0.007
1650 m	−0.78487	−0.61966	0.2578	0.013

Elevation was split each 50 meters into binary vectors with “0” encoding for communities sampled below a given threshold and “1” for communities sampled above or exactly at the threshold. Using the Vegan package⁸², we performed a non-metric multidimensional scaling (NMDS) of species presence/absence matrix. By fitting all binary vectors onto the ordination, we retrieved R² values and considered the threshold with the highest R² score as the most meaningful cutoff between lowland and highland. The analysis depicted two maxima with very similar R² values, 850 and 1250 meters above sea level (in bold).

Quick links

Search