Q1_1a:Sites from habitat A have a lower abundance of species 1 than sites from habitats B and C. Therefore, with repect to the abundance of species 1, habitats B and C are similar to one another and they are both different to habitat A.
Q1_1b:Sites from habitat A and C have a lower abundance of species 2 than sites from habitats B. Therefore, with repect to the abundance of species 2, habitats A and C are similar to one another and they are both different to habitat B.
Q1_1c:The abundance of species 5 appears to be highly variable and unrelated to the habitat types. On the basis of species 5 abundances, all habitats are similar and differences between sites is not related to habitats.
Q1_1d:Sites from habitat A have a higher abundance of species 8 than sites from habitats B and C. Therefore, with repect to the abundance of species 8, habitats B and C are similar to one another and they are both different to habitat A. This is the opposite trend to that of species 1. It appears that the abundances of species 1 and species 8 are negatively correlated.
Q1_2a:Species 1 and 8 are highly negatively correlated. There are also many mild correlations between species 1, 7, 3, and 4.
Q1_4a:Components 1, 2 and 3 explain a disproportionate amount of the original variation suggesting that these 3 new components (factors) explain the majority of the original variation.
Q1_4b:41.19158
Q1_4c:75.80812
Q1_4d:There is no obvious 'kink' in the scree plot, therefore use the 'greater than 1' rule in which only components that have eigenvalues greater than 1 are retained.
Q1_5a:Sites appear to cluster out according to habitat. That is the vegetation communities do appear to be different in the different habitats
Q1_6:Species 1,8,3 and 7 appear to have contributed most to principal component 1, species 2 and 4 contributed most to principal component 2 and species 5 contributed most to principal component 3.
Q2_1:Quadrats 1 and 2 are very similar, as are 3 adn 4 and 9 and 10
Q2_2a:Principal component 1 - 57.9% Principal component 2 - 21.1% Principal component 3 - 9.6%
Q2_2b:2
Q2_2c:There is a classic 'horse-shoe' pattern. Such a pattern is somewhat artificial and arrises due to asymptotic differences between site pairs that have little to nothing in common.
Q2_3:Yes
Q2_4:Environmental data could be correlated against each of the principal components. Each principal component is a measure of the vegetation community. They do not represent any one species, rather they represent patterns amoungs species and therefore may represent major processes (such as a response to a enviromental condition). Thus by correlating each of these principal components against some other environmental characteristic(s), it may be possible to identify important environmental characteristics that affect the vegetation communities.
Q3_1_1:5.916080
Q3_1_2:6.403124
Q3_1_3:3.000000
Q3_1_4:4.123106
Q3_1_5:0.52941176
Q3_1_6:1.414214
Q3_1_7:3.162278
Q3_1_8:9.273618
Q3_1_9:0.47368421
Q3_1_10:0.07692308
Q3_1_11:3.741657
Q3_1_12:9.486833
Q3_1_13:0.38461538
Q3_1_14:0.20000000
Q3_1_15:0.27272727
Q3_1_16:6.164414
Q3_1_17:0.71428571
Q3_1_18:0.85714286
Q3_1_19:0.75000000
Q3_1_20:0.80000000
Q3_1a:In general the two distance measures agree. Quadrats 2 and 3 are the most similar. Bray-Curtis suggests that quadrats 2 and 5 are the most similar followed by quadrats 4 and 5, however Euclidean distances suggest that quadrats 3 and 5 are the most similar followed by 2 and 5.
Q4_1a:As all observations are in the same units, no standardizations are necesary. Observations are measurements (cover abundance) of ecological units (species) and therefore Bray-Curtis distance matrix probably most appropriate.
Q4_1b:Different variables measured on different scales and some variables have larger values than others, therefore should standardise (scale) the variables. Because the observations are measurements of environmental characteristics, euclidean distances are adequate.
Q4_2a:Bray-Curtis dissimilarity matrix
Q4_2b:scaled Euclidean distance matrix
Q4_3a:0.4331
Q4_3b:<0.001
Q5_2a:0.2579
Q5_2b:Y
Q5_2c:The rank dissimilarities between groups was greater than than within groups for BF (biological farming) and HF (hobby farming), but not for NM (natural management) and SF (standard farming).
Q6_1a:0.0837
Q6_1b:There is an extremely good match between the patterns of site dissimilarities based on the original distance matrix and the new mds matrix
Q6_1c:Y
Q6_1d:y
Q6_1e:Latitude and longitude!
Q7_2a:13.85
Q7_2b:There is a reasonably good correspondance between the patterns amungst sites based on (distances) all 102 bird species abundances and the patterns amungst sites based on two new mds variables (distances).
Q7_2c:There is a non-linear relationship between bray-curtis distances and MDS distances, therefore non-metric MDS is best.
Q7_2d:The bird communities of some habitats appear to be distinct. Red gum, Gippsland manna gum and Box-ironbark communities appear to be distinct. Purhaps not surprisingly, bird communities from mixed forest sites are highly variable. This probably reflects the fact that these forest types may contain elements in common with a number of other forest types.
Q8_2a:
Q8_2b:
Q8_3b:
Q8_3b: