This book is part of the InfoQ Enterprise Software Development series of books. .... How does the team decide which stories to include in the sprint? ........27 ..... It may not be the way you do Scrum, it's how Henrik's team does Scrum. You may ask
The happiest people don't have the best of everything, they just make the best of everything. Anony
Idea Transcript
How do we measure diversity? EESC04 Afternoon Lab
What are the common ways of measuring diversity? • Richness: # of species • Relative abundance: proportion of a species in a community • Evenness: how well represented is a species?
Scale matters: Diversity at different levels • α-diversity – local diversity (single site) – Species Richness – Shannon-Wiener Index – Simpson Index
• β-diversity – change in diversity between sites – Sorensen Index (also known as Bray-Curtis) – Jaccard Index
• γ-diversity – regional diversity • ε-diversity if we are looking at a larger scale
Which site has the highest alpha-diversity?
Region X
Region Y
Site A
Site B
Site C
Site D
Which site has the highest alpha-diversity?
Region X
Region Y
Site A
Site B
Site C
Site D
Which region has the highest beta-diversity?
Region X
Region Y
Site A
Site B
Site C
Site D
Which region has the highest beta-diversity?
Region X
Region Y
Site A
Site B
Site C
Site D
Which region has the highest gamma-diversity?
Region X
Region Y
Site A
Site B
Site C
Site D
Which region has the highest gamma-diversity?
Region X
Region Y
Site A
Site B
Site C
Site D
Some useful indices • • • • •
Shannon-Wiener Index Shannon Evenness Simpson Index Margalef Index Berger-Parker Index
Shannon-Wiener Index • 𝐻′ = − 𝑆𝑖=1 𝑝𝑖 ln 𝑝𝑖 • pi = relative frequency of species # 𝑜𝑓 𝑖𝑛𝑑𝑖𝑣𝑖𝑑𝑢𝑎𝑙𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑠𝑝𝑒𝑐𝑖𝑒𝑠 𝑡𝑜𝑡𝑎𝑙 𝑖𝑛𝑑𝑖𝑣𝑖𝑑𝑢𝑎𝑙𝑠
• Values range from 0 – 5 – Usually between 1.5 and 3.5
• Sensitive to sampling bias
Shannon Evenness • 𝐸𝐻 =
𝐻′ 𝐻𝑚𝑎𝑥
• 𝐻𝑚𝑎𝑥 = ln 𝑆, where S = species richness – ie: 𝐸𝐻 =
𝐻′ ln 𝑆
• E ranges from 0 – 1 (1 is most even)
Simpson Index • 𝐷=
𝑆 𝑛𝑖 𝑛𝑖 −1 𝑖=1 𝑁(𝑁−1)
• ni = # of individuals of each species • N = total individuals • Less sensitive to species richness and weighs more abundant species more • Less sensitive than Shannon-Wiener
Margalef and Berger-Parker • Margalef: 𝑆−1 – 𝐷𝑀𝐺 = ln 𝑁 – Accounts for biases in sampling size and effort – S = species richness – N = total # of individuals
• Berger-Parker: 𝑁𝑚𝑎𝑥 – 𝐷𝐵𝑃 = 𝑁 – Looks at how important the most abundance species is – Nmax = # of individuals from most abundant species – Not very informative as it just looks at one species
Jaccard and Sorensen • Jaccard: – 𝑆𝐽 =
𝑎 𝑎+𝑏+𝑐
– a = # of species found in all sites – b = # of species found in site 1 – c = # of species found in site 2 • Comparison between 2 sites • Can’t use these for our data
• Sorensen: – 𝑆𝑆 =
2𝑎 2𝑎+𝑏+𝑐
– a = # of species found in all sites – b = # of species found in sites 1 – c = # of species found in sites 2