Niches and Areas of Distribution

Niches and Areas of Distribution

Photo by Warren Apel Niches and Areas of Distribution I Jorge Soberon Museum of Natural History, University of Kansas Caveat Emptor

This is a talk about ideas being developed as we speak. Nothing is yet firmly established We encourage dissent and questioning Key papers: Pulliam, 2000, Ecol. Lett. 3:349-361 Pearson & Dawson, 2003, Glob. Ecol. Biogeog. 12:361-371 Araujo & Guisan, 2005, Glob. Ecol. Biogeog. 33:1677-1688

Kearny 2006, Oikos 115: 186-191 Soberon & Peterson, 2005, Biodiv. Info. 2:1-10 Soberon 2007, Ecol. Lett. 10:1115-1123 Purpose: to clarify very muddled terminology We need concepts of area of distribution of a species which are clear and operational. The same for niche. Relate the concepts of areas of distribution and niche. Around 1990 three things happened 1. Large databases of

presences of species (mainly computerized scientific collections) began being accessible at significant amounts II. GIS Geographical Information Systems technology became widely accessible to ecologists and biogeographers III. Niche Modeling

Since the 1980s, the Australians began doing bioclimatic modeling for applied entomology purposes (Climex) Later they developed Bioclim, DOMAIN, and then, GARP Today, there are about 15 methods to do niche modeling, most of them in freeware form Using those tools Since about 1995 there has been an explosion in the application of ecological niche modeling methods to estimate areas of distribution,

potential or actual. The methods are used to predict invasive-species routes, spread of vectors of diseases, restoration ecology, conservation, and many topics in fundamental ecology and biogeography. Literally hundreds of papers have been published, about the potential and/or actual distributions of thousands of species And yet Geographic distributional areas are the shadows produced by taxa on the geographical screen. To study them one needs to measure ghosts A. Rapoport. Areography: Geographical Strategies of Species, 1982 Most ecologists would agree that niche is a central

concept of ecology, even though we do not know exactly what it means L. Real and S. Levin, Theoretical advances: the role of theory in the rise of modern ecology 1991 In other words: We (the niche modelers) are doing wholesale estimation of those shadows (geographical areas) by using a nonconcept (niche)! Not nice We need to define ENM terms and methodology

Preferably operational definitions Relevant in the field of modeling distribution areas Preferably mathematical (formal) Avoid breaking with tradition, if possible Let us begin with area. In books they appear as little maps Some theorists even treat them as if they were physical entities. But areas are composed by the presence (detected) of individuals 2

1 0 -1 -2 -2 -1 0 1 Jim Brown suggested a Gedankenexperiment

2 2 1 Imagine that we make individuals fluoresce... 0 -1 -2 -2 -1

0 1 2 2 1 0 -1 -2 -2

-1 0 1 2 We see that to define area of distribution we need to: Agree on what it is meant by a presence (sink or source populations). Potential or actual. Agree at what resolutions makes sense to talk about areas of distribution?

Normally, this would be low resolutions > 103 km2 Inside the cells, we are in the domain of ecology, outside, it is biogeography or macroecology So areas are... Regions of the world defined by the detection of individuals within cells in grids There are several things that we can detect: reproductive individuals, just individuals, regions with positive growth rates... Although a lot of those can be defined at high resolutions, areas are defined at low resolutions. It is still a dream to measure the movements of all the individuals of a species...

What determines where do we find individuals? Autoecology, interactions, migration patterns, historical factors, operating with different strenghts at different spatiotemporal scales. There are equations (horribly difficult) describing this. Species j j 1 dxi j j j j ri (e ) i ( x ; Ri ) ( x; T) j x i dt

Grid cell j 1 dxi j j j j j ri (e ) i ( x ; Ri ) ( x; T) j x i dt

j j j i ( x ; Ri ) j ri (e ) Grinnell, scenopoetic B A Elton, bionomic

( x ; T) j M BAM Diagram G Movements So we can define stuff We can define, rigorously, areas of distributions in terms of whether the xij (t) are positive, or above

a threshold, or whether the intrinsic growth rate is positive (potential area of distribution) J F { j G | r (e j ) 0} A J C { j G | J O { j | dx j dt 0} A B x j 0

dx j dt 0 x j 0 and x j (t ) 0, t0 t t1} A B M And niches? Clearly, the spatial regions determined by the parameters in those horrible equations have environments. Niches may be

defined using the environmental variables in the occupied cells. A useful and forgotten distinction of Hutchinson (1978) Scenopoetic variables. Non-interacting, slowly changing from a species point of view. Bionomic variables. Coupled, fast changing Hutchinsons nice distinction was quickly forgotten and then reinvented by Austin,

Begon, and others. Niches can be defined using the same equations that govern area Fundamental niche Intrinsic Growth Rate (Scenopoetic) j Resource-consumer dynamics, competitors, predator-prey (bionomic). Migration colonization, history 1 dxi

j j j j j ri (e ) i ( x ; R i ) ( x; T) j x i dt So, the niches are the sets of environments present in the different areas: An abiotic niche: the set of environments in the area where the intrinsic growth rate is positive A colonizable niche: the set of environments in the

area where both biotic and abiotic conditions are right A occupied niche, the set of the environments in the area where the species actually occurs (J F ) N F (J C ) NC (J O ) N O Hutchinsons terminology 1. Fundamental niche. Defined without reference to scenopetic

and biotic variables. 2. Realized niche. Fundamental reduced by competition. The entire biotope occupied 1. Abiotic niche. Defined using only scenopoetic variables. 2. Colonizable niche. The abiotic reduced by competition. 3. Occupied niche. The realized niche taking into account movement restrictions Some areas and niches

Abiotic A ( A) N F 1 (N A ) A J A B JO JO Colonizable and occupied

J O JO J C M N O N C Occupied = colonizable? (J O JO ) (J O ) (JO ) N O NO N C The Eltonian Noise Hypothesis I. At the scale at what we work (cells > 100 km2), the biotic factors become much less relevant than the

abiotic ones: the signal is dominated by abiotic factors, and the biotic factors act as noise B A JO JO M Eltonian noise hypothesis II. The Eltonian processes are of very high

resolution In the large cells at which scenopoetic variables are measured competitive exclusion may take place without affecting the totality of populations. B M A Examples and counterexamples Cactoblastis cactorum Obligate pollinators

Extremely Important I. By assuming non-interactivity we can represent the environments (and therefore the Grinnell niches) by sets of numbers. The interactive and dynamic Elton niches must be represented by the parameters of functions (sometimes by the isoclines), because one needs to measure both the use of a resource and its impact (see Chase and Leibold, 2003, Ecological Niches) Extremely Important II. Since we are defining our Grinnell niches using such scenopoetic variables, it means that we can use: Widely available datasets At low resolutions

Towards concluding So it is useful to define areas as subsets of cells in a geographic grid, defined by biological properties of species (presence, presence of viable populations, suitable conditions for survivorship,) And Grinnell niches as the subsets of scenopoetic variables associated to those areas And of course, Grinnell niches are multidimensional, but in the space of scenopoetic variables The concept of here defined as the sum of all the

environmental factors acting on an organism; the niche thus defined is a region of an ndimensional hyperspace Multidimensionality applies to both scales of niches In Grinnells (related to distribution, geographic), variables are elevation, orientation, geology, climate... These variables are less interactive

In Eltons (more trophic, ecological), variables may be nutrient concentrations, food size distributions, prey, predators, competitors and mutualist densities... These variables are highly interactive So, it is useful to accept tow levels Grinnell Niches, low resolution, scenopoetic variables. Elton Niches, high resolution, bionomic

variables. They may be the cause of reduction from Fundamental to Realized M B A This distinction is based on: Accepting the assumption that differences in certain extreme types of variables: non-interactive, meaningful at low resolutions, vs interactive, meaningful at high resolutions do exist. How valid is this simplification?

The fact that ENM is predictive is evidence in favor. We need to establish the scope of validity of the assumption. At high resolutions probably it will break down. Thanks to Town Peterson, Richard Pearson, Enrique Martinez, Miguel Nakamura, Carlos Martinez del Rio, the students of the Niche Seminar Photo by Warren Apel Niches and

Areas of Distribution II Jorge Soberon Museum of Natural History, University of Kansas So how we measure a Grinnell Niche? 1. We go and see where it is that a species occurs (without immigration) 2. We measure environmental variables in those places. 3. We extrapolate from a few points to an area. 4. We obtain the corresponding environmental

values. 5. This is the niche. There are Several Ways of Calculating the Niche Physical variable 2 Bioclim Clusters About 15 others... Physical variable 1 Bivariate Histogram E-space. Many variables.

Scenopoetic. Low resolution. Low precision 5 10 25 50 75 100 125 150 175 200 225 250 275 above

G-space. Two dimensions Low resolution (pixels of 1 km2) Notice: Conceptually, every algorithm tries to recover the realized Grinnell niche, since the data points come from the actual distribution JO NO JO JO 1. However, when we map back from E to G, in general, we do not recover JO since there may be other spatial regions with the same niche. We recover JC. In general, one has to deal with this #

S # S # S # S # S # S # #S

S S# # S # S # S # # S S # S

# S # S # S # S # S # #S # S S

# S # S # S # S # S Tamandua mexicana #

S # S # S #S ### # S S # # S S

#SS S # S # S #S S # # #S S # #S S S # # #S

# S # S S # S # S # S # S When -1(NO) is different from JO? 1

J1 E1 1 J2 G-space E-space Begin with presence (absence) data Using Garp, Maxent, GAM, estimate the Occupied Niche

With GIS display the geographical extent of the Occupied Niche If you want the Jo (rather than JC), then clip the geographical extent of the ER using biological connectivity or historical reasoning. A real-life example: Baronia brevicornis and its single food plant, Acacia cochliacantha For this example the following layers were used, all at 1:1,000,000:

Aspect Elevation Mean Temperature Average minimum daily temperature Average maximum daily temperature Yearly precipitation MME about 500 meters B. brevicornis Abiotic Niche using BS Garp II: Estimating the Area of Accesibility

From where? What is the initial condition? At what scale? In relation to what vagility parameters? At certain scales, one can assume that biogeography is a good surrogate for the accesibility areas, this is, we assume that if a species is present in a given biogeographical region, it can reach all of it. B. brevicornis Biogeographical Provinces B. brevicornis. Biogeographical Provinces as Surrogates for Accesibility Areas B. Brevicornis Niche Accesibility

III. Estimating an Obligate Interaction B. brevicornis only food plant is the legume Acacia cochliacantha A similar analysis is the repeated for A. cochliacantha A. cochliacantha Abiotic Niche A. cochliacantha accesibility region (biogeography) A. cochliacantha niche accesibility B. brevicornis (red) A. cochliacantha (blue) Finally, is it the occupied or the abiotic niche?

A J O B ( A) N A N O (J O ) A JO B A JO M

A JO ( A) N A N O (J O ) M Conclusion. We still have a lot to learn, for example about: The process of reduction from Abiotic to Occupied, by Eltonian factors. This is a problem in scaling The basic assumption of using ENM to study invasions is that NO = NC.Is this true? When? A more standardized procedure to obtain different Areas and Grinnellian Niches using ENM Developing validation techniques based on different configurations of the BAM diagram

And many other things Acknowledgments Providers of museum data, for the huge effort they have put in collecting, curating and identifying specimens, and then sharing their results The owners of the land where those specimens were originally taken Very Fundamental Concepts: To each region in G-space there are corresponding ones in E-space (G) = E There is an

inverse operation that maps from G to E: (E) = G G-Space, 2 dimensions E-Space n dimensions More Fundamental Concepts. Niches (Grinnell) are subsets of spaces of environmental variables corresponding to subsets of geographical space where a species live, or can live.

The function that maps from G to E is defined very simply by the tables in the GIS that link coordinates with environmental values. And viceversa, for the relation -1 that maps back E to G Gedankenexperiment I: We know all regions where intrinsic growth rate is positive (this is, we disregard biotic interactions and movement limitations) GF We agree to call the corresponding Environmental subset the Fundamental Niche, EF = (GF) GF It ought to be true that -1(EF)= -1((GF))=GF

Gedankenexperiment II: We know now the area of distribution of a species GR We agree to call the corresponding Environmental subset the Realized Niche, ER = (GR) GR Now it is obvious that -1(ER) = may not be = GR Tamandua mexicana 140 180 220

20 40 60 80 120 260 100 100 140 180 220 180 220 Bio1 4000

7000 TempRng 100 1000 AnnPrecip 20 60 CoefVarPrecip 180

220 260 1000 4000 7000 # S # S

# S # S # S # S # #S S S# # S #

S # S # # S S # S # S # S

# S # S # S # #S # S S # S # S

# S # S # S # S # S # S

#S ### # S S # # S S #SS S # S # S #S

S # # #S S # #S S S # # #S # S # S S # S

# S # S # S When -1(ER) is different from GR 1 G1 E1 1

G2 G-space E-space ?

Recently Viewed Presentations

  • COP330 Object Oriented Programming in C++ Fall 2008

    COP330 Object Oriented Programming in C++ Fall 2008

    Programming in C++ Fall 2008 Dr. David A. Gaitros [email protected] Class Access Functions One of the important aspects of OOP is information hiding and protection. We do not let users of the classes have direct access to either look at...
  • Week 1, Nature, the maestro

    Week 1, Nature, the maestro

    It left a crater of . radioactive glass. in the desert 10 feet (3 m) deep and 1,100 feet (330 m) wide. Trinitite The name given to the radioactive glass substance formed from super-heated sand in the blast crater during...
  • The Jews 1940 - 1945

    The Jews 1940 - 1945

    The Jews living in these Ghettos were to be used as a cheap source of labour. Conditions in the Ghettos were designed to be so bad that many die whilst the rest would be willing to leave these areas in...
  • EMC / VMware Cloud Strategy Answering "Why EMC for VMware?"

    EMC / VMware Cloud Strategy Answering "Why EMC for VMware?"

    That's where EMC's backup and recovery portfolio provides the right framework to improve your backup and recovery architecture, regardless of where you are on this journey. Let's take a quick look at Data Domain, Avamar, Disk Library for mainframe, NetWorker...
  • POSTER MAKER Save Lives by Supporting the 5th

    POSTER MAKER Save Lives by Supporting the 5th

    Every year health care-associated infections affect hundreds of millions of patients around the globe; people die or are considerably affected by severe infections acquired during patient care, including many that are resistant to antibiotics. Most of infections can be avoidable....
  • Proposal for Advisory Services to DA Healthcare Hospital

    Proposal for Advisory Services to DA Healthcare Hospital

    Port project, Refinery project Financing in Oman. ... Divestment of stake in a baby food manufacturing company. ... In-depth Economic Analysis on the Status of Roller Flour Mill Industry in Tamil Nadu (India) with a major emphasis on the demand...
  • Interference & Standing Waves - Uplift Education

    Interference & Standing Waves - Uplift Education

    Destructive interference. Occurs when the two waves oscillate the particles of the medium in opposite directions at a given point in space. Causes a decrease in amplitude (and thus energy) at that point. Causes a decrease (or absence!) of light,...
  • Finding Information on the Internet  Its as easy

    Finding Information on the Internet Its as easy

    Finding Information on the Internet Subscription databases Subject directories Search engines 3. Search Engines Finding Information on the Internet It's as easy as Finding Information on the Internet Subscription databases Subject directories Search engines 3. Search Engines * * OK,...