Excavating the Bare Bones of Community Structure Using Loop Analysis - PowerPoint Presentation

Slide1

Excavating the Bare Bones of Community Structure Using Loop Analysis

By

Patricia Lane

Dalhousie UniversitySlide2

Goal

: to illustrate how Dick Levins’ loop analysis is useful for analyzing complex systems using a marine ecosystem example

Rationale: we have one ocean, which is constantly under threat; we need to understand how perturbations affect ecological networks through a myriad of pathways and feedbacks

Although this is a marine example, a similar use of loop analysis could be helpful to understand any complex system with a detailed data set.Slide3

Traditional DescriptiveFood Webs

This diagram is essentially a summary of every feeding relationship known to the author

Certainly

not stable-excessive connectionsPoor sense of functional groups and links that are ecologically relevantCannot locate parameter

inputs

Not

useful for making predictions about perturbationsSlide4

Objectives

To introduce the field and laboratory studies at the Marine Ecosystem Research Laboratory (MERL)

(NOAA Grant to Levins/Lane)

To illustrate ecological skeletons and some of their propertiesTo acknowledge Dick’s considerable influence on this and other workSlide5

1. Methods: Field Studies

10 cruises in Narragansett Bay

Temperature/lightNutrients & water quality parametersPhytoplanktonZooplanktonSlide6

1.

Methods:

Laboratory Studies Nutrient Enrichment ‘Press’ Experiment at MERLTanks: 1.8 m diameter

5.5 m deep with sediments3 controls

6 treatments

1X, 2X, 4X, 8X, 16X, 32X

~sewage effluents = 1X

12.8N, 1.0P,

0.9Si

-inorganic

Simulate Bay mixing, light, te

mperature, residence time,

b

enthic-pelagic coupling

Slide7

1. Methods: Loop Analysis Models

Loop Analysis

Data Preparation

Prepare a qualitative correlation matrixDetermine functional groups and loop variablesCalculate directed changes in data sets by loop variable

Prepare models

Fit loop predictions to data –try to achieve 90-100% accuracy

Created by Dick

Levins

in 1973 -NY Acad. Science paper

Signed digraphs, qualitative analysis that predicts changes in standing stocks of variables and their turnover rates following a parameter change

Also gives measures of network structure and stability

Used theoretically to explore complex systems as well as with data

setsSlide8

Loop Variables (~300 taxonic

units)

Si=silicaN

=nitrogen/phosphorusN2 =organic nitrogenA1=diatomsA2=dinoflagellatesA3=luxury consuming diatoms

A

4

=silica flagellates

A5

=

microflagellates

/monads

Z

1

& Z

3

= copepods

Z2=

copepodids

and

nauplii

Z4 =

cladocerans

M=mollusc larvae

C=

cirriped

larvae

P=polychaete larvae

G=

gammarids

R=rotifers

S = Sagitta elegansD = decapodsMD=medusaeSlide9

2. Results: 9 Field Loop Models (

Annual Cycle: No 2 models identical

)Slide10

Ecological Skeleton (ES) of the Field Loops

[- +] with 9 link types

Variables in a minimum of 3 models

Solid links 2/3 of networksDashed Links

≥ 1/3 of networks

63 feedback loops & 4400 pathways

ES does not appear in nature-too connected, but it is a framework for the community

~ Chemical isomers

 Slide11

Core Loops = Ecological Skeleton (ES)

LA provides basic structure of an ecosystem (or any complex system) as the prevalent variables, links and parameter inputs that explain the pattern of changes in standing stocks

Thus, the Ecological Skeleton (ES) is a form of minimal description (bare bones) used

to capture the main aspects of community structure & how parts and whole interact Most ‘nibbling’ in traditional webs is irrelevant – ES helps get rid of the ‘noisy’ detailSlide12

Some Loop Analysis Benefits

Marine food webs

have a 3 tier structure, largely composed of L1 and L2 feedbacks, short negative feedback loops that are stabilizing, but several

links have 2 or more link types = volatile links (often nonlinearities can be identified)Can identify missing variables (N3 and Ax) and link types (silica-diatoms) that are counter-intuitive, but later explained by biologyCan identify where parameter inputs enter-not always the manipulated oneCan observe how the variables and links change through an annual cycle and how one

community prefigures

another

Loop analysis helps identify what is necessary to measure with great economy of resources (10-15%)

Can compare over ecosystems to measure ES similarities

IN

SUMMARY, LA PROVIDES NEW INFORMATION AND

INSIGHTSSlide13

Comparison of the Field and Lab ESs

Narragansett

Bay(9 cruise dates)

MERL Tanks(135 dates x tanks)Slide14

Some Results

Field

3 large pelagic predators out-competed in MERL tanksPredictions 145 (3 wrong or 2% and 6 zeroes or 4%) 94% correctSystem of small plankton less discrete than in tanksConnectance

: 16%-in line with theoretical predictions of Gardner and Ashby of 13 +/-2 % dividing stable and unstable networksMERL Lab (X-32X nutrient additions)Benthos absorbed most of increased productivity-strong coupling

1046 predictions (12 wrong or 1% and 102 zeroes or 10%) 88% correct

Little response of small algae and copepods up trophic

gradient

Connectance

: higher (19%)-perhaps because of confined space

Limits to enrichment after 4XSlide15

Structural Complexity Calculations

MACRO-UNIVERSE: A

20 variable loop model has 3

N² or 3400 or 7.06 x 10

190

mathematically possible configurations-but most are not biologically reasonable

Lab: 15 variables = 3

225

or 2.25 x 10

107

MICRO-UNIVERSE: 6 x 10

7

to 1.3 x 10

12

for the tanks-biologically reasonable

THE TRUTH “truly” IS THE

WHOLE in the structural and functional complexity of all of these feedback relationships in ecosystems

Stable UnstableSlide16

Ecological Complexity

Dick has called complexity, “the central intellectual problem of our

time.”A complex system cannot be completely described: has at least on incomputable model. (Robert Rosen)Both structural and functional definitions

To advance ecological theory, we need to understand the level of complexity in ecological skeletons with their rich feedbacks, only studying pairwise interactions in natural communities is not enough.Need to develop ecological skeletons for all major kinds of ecosystems Slide17

Emergence

Core models or ecological skeletons (ES) have been computed as summaries of a set of loop models and therefore are calculated as collective properties

Over a number of marine ecosystems, the ES are robust and very similar, transcending space, time, taxonomy, field & lab methods and investigator intuition: EMERGENT???ESs for freshwater ecosystems are different and less robust and have more structural change with nutrient enrichmentAttempts to ‘break out’ of the marine ES have been unsuccessful

Can’t prove one ES is the best or most optimum descriptorSlide18

4. Thank You Rosario

When the MERL work first

started, Dick had a massive heart attack and many of us wondered if he would reach his 55th birthday let alone his 85th.

Although Rosario is not here today, she has to be given a lot of credit.She oversaw Dick’s change of lifestyle, which gave him another 30+ years of great accomplishments, and he gave us innumerable academic gifts, friendships, and life-altering influences.Slide19

4. Thank You Dick

Teaching

-Communities and Ecosystems-Political Ecology -

Sustainability & Global Change 2. Research-Tropical fieldwork & theoretical ecology/Ph.D.-Ecological networks/loop analysis

Political

Work

-

International Development-with a political conscience

-

Cuban

Model-an

alternative to neoliberal

globalization

-H

ow

Cuba could be the First Sustainable

Society?Slide20

Happy

85th

Birthday to an amazing complex system with delightful emergent properties & unique s

elf-organization always exuding creativity and humanity in equal portions.