Ti

Sa

b

a

KHBSDRE

1

disohobgcthlca

w

s

1h

Ecological Indicators 28 (2013) 48–53

Contents lists available at SciVerse ScienceDirect

Ecological Indicators

jo ur n al homep ag e: www.elsev ier .com/ locate /eco l ind

he properties of the ecological hierarchy and their application as ecologicalndicators

.E. Jørgensena,∗, S.N. Nielsenb

Copenhagen University, Institute A, Section of Environmental Chemistry, University Park 2, Copenhagen Ø, DenmarkECO-Soft, Kålagervej 16, DK-2300 Copenhagen S, Denmark

r t i c l e i n f o

eywords:ierarchical organizationiodiversitypecies diversityamping effectandom disturbancescosystem stability

a b s t r a c t

After a short overview of the hierarchical organization, that characterizes ecological systems, a quantifi-cation of the openness of the various hierarchical levels is introduced. By the use of statistical calculations,it is furthermore shown that the random variations due to environmental disturbances in one level ofthe hierarchy are averaged and therefore result in less variations or disturbances in the next level. Ran-dom disturbances are with other words damped when we go up through the hierarchical levels, whichobviously is a clear advantage by the hierarchical organization. Thus it is shown that diversity at all levelshas the consequences of ensuring a more stable system and less sensitive to environmental disturbances.This is a result that is in contradiction to earlier findings of May (1973, 1981). Approximate calculationsof this damping effect can be carried out. The level out effects of the hierarchical organization and the

recovery time makes it possible for ecosystems to cope with the relationship between the frequency andthe magnitude of the disturbances. These important properties that are crucial for the reactions of thevarious hierarchical levels to the impacts on ecosystems, are applied in a discussion of the choice of eco-logical indicators and the applicability of these indicators are demonstrated. The presented hierarchicalproperties entail that biodiversity on all hierarchical levels is a very important ecological indicator, whichis the core topic for the discussion, that summarizes the results and conclusions.

. Introduction: the hierarchical organization

The ecosphere is organized hierarchically and may be brieflyescribed as going from molecules, via cells, tissues, organs,

ndividuals, species, populations, communities, ecosystems, land-capes, regions to the ecosphere. The biological hierarchy is easy tobserve. The biochemical processes take place in the cells, whichave molecular components and structure via the genomes notnly to control the processes, but also to protect the genome itselfy membranes. In vertebrates, there are many different types androups of cells, which are generally determined and specialized toarry out the specific biochemical processes that are characteris-ic for different organs: the liver, the muscles, the kidneys and theeart. The cells that carry out the processes that take place in the

iver make up the liver and so on. It is a proper and effective hierar-hical solution that the cells that have certain biochemical functions

re working together to ensure the functions of the organs.

The next hierarchical level above the organs is the organismhich may be represented by different species. The individual

∗ Corresponding author.E-mail addresses: msijapan@hotmail.com (S.E. Jørgensen),

oerennorsnielsen@gmail.com (S.N. Nielsen).

470-160X/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.ecolind.2012.04.010

© 2012 Elsevier Ltd. All rights reserved.

organisms are working together in populations, that apply numer-ous methods to ensure survival and growth for the individuals andfor the entire population. The grazers form a herd that makes itmore difficult for the predators to attack the individuals of the herd,including a protection of the relatively newborn offsprings. On theother side, the predators hunt together to obtain by cooperationa higher probability for a successful hunting. The individuals ofpopulations are also using many different forms of communicationamong them in order to increase the probability for survival. Popu-lations are interacting in a network and make up together with thenon-biological components of the environment, the ecosystem. Theinteractions in networks have (Patten, 1991, 1992; Jørgensen et al.,2007), a synergistic effect that is able to increase the utilization effi-ciency of matter, energy and information. Landscapes are formedby interactions among several ecosystems and regions compriseof many landscapes. The entire living matter, the landscapes andregions on the Earth make up the biosphere and the biosphere plusthe non-biological components are denoted the ecosphere.

A more complete version of the biological–ecological hierar-chy is illustrated in Fig. 1. The figure indicates the corresponding

approximate spatial scale of the respective compartments. Theorganelle, a functional grouping of bio-molecules, and the tissue,a functional grouping of cells, may be included as levels betweenmolecules and cells and between cells and organs.

S.E. Jørgensen, S.N. Nielsen / Ecological Indicators 28 (2013) 48–53 49

Hiera rchical levels

Atoms Cell s Organs Spec ies Popu lation s Ecosystems Landsc ape Regions Th e ecosphere

-10 - 6 - 2 + 2 + 4 +6 +7 +8

cated

ierarc

hfpottea

aacta

omsoHiepot

di

vhawcblabmd

TRe

ic

Approx imat e log (linear scale in meter) is ind i

Fig. 1. The presented h

Following Simon (1973), the hierarchical approach represents aeuristic supposition to better understand complex systems, and

ollowing Nielsen and Müller (2000) hierarchical approaches arererequisites for the definition of emergent properties in self-rganized systems and to the possibilities to apply informationheory and quantify the information gained by organizing the sys-ems in a hierarchical manner (Nielsen, 2000a). In facts the wholecosystem may be redefined as one larger series of embedded hier-rchical levels (Nielsen, 2000b).

Hierarchy theory was first formulated by Simon (1973), Allennd Starr (1982) and O’Neill et al. (1986). Kay (1984) has proposedn integrative concept of ecosystem based classification and con-eption, which is compatible with most of the existing approacheso ecological system analysis. Recently, there have been severalpplications in ecosystem analysis and landscape ecology.

Exchange of matter and information with the environment ofpen systems is in principle not absolutely necessary from a ther-odynamic point of view, as an energy input (non-isolation) is

ufficient. The system must be non-isolated to ensure an inflowf energy for the maintenance of the system far from equilibrium.owever, ecosystems are open, i.e. can also exchange matter and

nformation with the surroundings, and this openness gives thecosystem additional advantages, by the input of chemical com-ounds needed for certain biological processes or by immigrationf species offering new possibilities for a more ordered structure ofhe system.

Higher levels change more slowly than lower levels due to theifference in spatial scale and the dynamics (how quickly the level

s able to change); see Table 1It will be discussed below quantitatively, how the random

ariations due to environmental disturbances in one level of theierarchy are averaged and why they therefore result in less vari-tions or disturbances in the next level. Random disturbances areith other words damped when we go up through the hierarchi-

al levels. It is a clear advantage by the hierarchical organization,ecause the over-all effect of disturbances is increasing with the

evel. Thus, the variations and disturbances of the lower levels have

dynamic that may affect locally the lower levels in the hierarchy,ut due to the hierarchical organization, the higher levels will beuch less affected because the variations caused by the random

isturbances are level out. Section 2 presents a quantification of

able 1elationship between hierarchical level, openness (area/volume ratio), and approximanergy/volume, space scale, time scale, and behavioral frequency.

Hierarchical level Opennessa,c (A/V, m−1) Energyb (kJ/m3)

Molecules 109 109

Cells 105 105

Organs 102 102

Organisms 1 1

Populations 10−2 10−2

Ecosystems 10−4 10−4

Ecosphere 10−7 10−7

a Openness, spatial scale and time scale are inverse related to hierarchical scale.b Energy and matter exchange at each level depend on openness, measured as available

n small packages (quanta, h�, where h is Planck’s constant and � is frequency), which makoupling makes energy usable at all hierarchical levels.

c Openness correlates with (and determines) the behavioral frequencies of hierarchica

. Notice the ax is is not l inear

hical levels are shown.

the “level off effect” of the hierarchical organization, followed bySection 3, that shows how the hierarchical organization and thedynamics on each level (see Table 1) are well fitted to the magnitudeof the environmental disturbances as function of the frequency. Themagnitudes of the environmental disturbances as function of thefrequency follow a power law (Bak, 1996). The recovery time makesit possible for ecosystems to cope with the relationship between thefrequency and the magnitude of the disturbances. These importantproperties that are crucial for the reactions of the various hierar-chical levels to the impacts on ecosystems, should obviously bereflected in the selection of indicators. Section 4 proposes such eco-logical indicators inspired by the presented hierarchical propertiesand a few examples will illustrate the applicability of these indica-tors. The presented hierarchical properties entail that biodiversityis a very important ecological indicator, which is in disagreementwith (May, 1973). This topic is discussed in Section 4. The last sec-tion is devoted to a summary and the conclusions.

2. Interactions between the hierarchical levels

Fig. 2 shows how the processes on one level of the hierarchydetermine the conditions in the next level immediately above andhow a level regulates and controls a lower level by feedbacks. Fora strict thermodynamic interpretation of the relations in such sys-tems see Nielsen (2009). A specific level consists of interacting andcooperative entities that are again in turn integrated as componentsin a higher organizational level. The interaction of the entities inone level produce an integral activity of the whole, or expresseddifferently the dynamic of a lower level generate the behavior ofthe higher level andd vice versa (Nielsen, 2009; Ulanowicz, 2009).The variation of a whole level is significantly smaller than the sumsof the variation of the parts, but the degree of freedom of sin-gle processes are limited by the feedback regulation (downwardcausation) and constraints from the higher level.

The random standard deviation of an average of n componentsis the average deviation of the components divided by n0.5. Thecells contain 108–1010 molecules. Cells have therefore a random

deviation caused for instance by environmental disturbances cor-responding to a deviation, that is about 10,000–100,000 timessmaller than the average molecular variations. The variation of theenergy content for instance caused by randomly environmental

te values of the four scale-hierarchical properties, presented by Simon (1973):

Space scalea (m) Time scalea (s) Dynamicsc (g/m3 s)

10−9 <10−3 104–106

10−5 10–103 1–102

10−2 104–106 10−3 to 0.11 106–108 10−5 to 10−3

102 108–1010 10−7 to 10−5

104 1010–1012 10−9 to 10−7

2 × 107 1013–1014 10−11 to 10−12

exchange area relative to volume. Electromagnetic energy as solar photons comeses only utilization at the molecular level possible. However, cross-scale interactive

l levels.

50 S.E. Jørgensen, S.N. Nielsen / Ecological Indicators 28 (2013) 48–53

Level of spec ies

Organ level

Cell le vel

Molecu lar level

The cell co ntrols

The bioc hemist ry

The bioc hemistry

determines the c ell

function s

The orga n controls

The final bioche mical

results of the cells

The integrate d cell

processes determine the

fun ction ality of the orga n

The organi sm deter-

mines the function a-

lity of the orga n

The pop ulation

determines the

framework for the

individuals

The individu als and th eir

interac tio ns det ermin e the

proper ties of the popul atio n

The su m of th e organ functions

determines the l ife con ditions

for the sp eci es

esente

dctsc

lattwiitvfhtctboTsccfe

R

wvc

nisb

Fig. 2. The four lower levels of the ecological hierarchy is pr

isturbances of molecules, may be ±10% of the average energyontent or ±0.1 × 109 = 108 kJ/m3 (see Table 1). The variation (dis-urbance) of the energy content of the cells is therefore 10,000 timesmaller or 10,000 kJ/m3, which is 10% of the energy content of theells. For all the levels of the hierarchy, it is similar.

The number of components which are contained in the nextevel of the hiearchy ensures that the variations of intensive vari-bles in each level are the same expressed as percentage, providedhat the focus is on intensive variables as the energy content andhe dynamics (see Table 1). For extensive variables, it is different,hich can best be illustrated by an example. If the ability of organ-

sms to find food due to random, environmental disturbances fornstance changes 100 kJ out of a total need of 1000 kJ or 10%, thenhe variation for an average organism in a herd with 10,000 indi-iduals to find food is 100 times less or 1 kJ or 0.1% of the valueor an organism., which means that the disturbance for the entireerd will be the average variation × the number of individuals inhe herd = 1 × 10,000 kJ = 10,000 kJ. If the variation would have beenarried over fully to the next level of the hierarchy unchangedhe variation of the ability to find food for the herd would haveeen 100 kJ × 10,000 = 1000 MJ or 100 times bigger. Table 2 (basedn Jørgensen, 2012) gives for the hierarchical levels presented inable 1, an overview of the approximate damping effect. It is pre-umed that a variation on the molecular level is 100% and theorresponding variations on each of the other six levels are indi-ated in the table as percentage for an extensive variable. It is foundor the average component in the (n + 1)th level from the followingquation:

= 1p0.5

here R is the ratio of the variation of the (n + 1)th level and theariation at the nth level and p = number of nth level componentsontained in the (n + 1)th level.

It is presumed that the number of components contained in the

ext hieararchical level is (the ratio of the two scales in m2), which

s in accordance with the allometric principles (Peters, 1983). Ithould, however, be considered a coarse approximation. This num-er is also indicated in the table. Moreover, the table shows the

d. The interactions between the levels have been indicated.

numerical total variation of the seven levels assuming that themolecular level has a variation of 100. It is indicated in the tableas variation relative, because it is the numerical value of the totalvariation of all the components in a considered level relative to avalue of 100 on the molecular level. It can be found by the followingequation:

TVn+1 = CVn+1 × p × TVn

100

TVn+1 = total relative variation for level n + 1; p = the number of com-ponents; CVn+1 = the variation of the average component of level(n + 1) in %; TVn = the total relative variation for level n.

Notice that the relative numeric total variation (disturbance, rel-ative to the disturbance at the molecular level) is increasing fromthe molecular level to the cell level but decreasing from the cells tothe ecosphere. The table shows clearly how the hierarchical orga-nization is damping variations or possibly random (environmental)disturbances. The consequence is that ecosystems and particularlythe ecosphere are very stable or expressed differently are very goodsurvivors. Usually – at least before the massive impact by man –ecosystems survive hundred thousands of years and the ecospherehas survived almost 4 billion years as there has been life non-stopon the Earth since the first primitive cells emerged 3.8 billion yearsago.

3. The frequency of disturbances

The magnitude of variations, disturbances and catastrophicevents has a frequency that varies according to the power law (seeBak, 1996). The frequency, F, as function of the magnitude, M, istherefore expressed by the following equation

F = a ∗ Mb (1)

An example is the distribution of earthquake intensities. Fig. 3shows as an example the frequency versus the earthquake inten-sities (M) in the Richter scale (which is logarithmic) in a district in

S.E. Jørgensen, S.N. Nielsen / Ecological Indicators 28 (2013) 48–53 51

Table 2Relationship between hierarchical level, openness (area/volume ratio), and approximate values of variations or disturbances, provided that the variation on the molecularlevel is 100%, space scale (taken from Table 1, the number contained in the next level and the total relative variation).

Hierarchical level Opennessa,c (A/V, m−1) Variationb (%) Space scalea (m) Number inc next level Value relatived

Molecules 109 100% 10−9 108 100Cells 105 0.01% 10−5 106 104

Organs 102 10−5% 10−2 104 10Organisms 1 10−7% 1 104 10−4

Populations 10−2 10−9% 102 104 10−11

Ecosystems 10−4 10−11% 104 4 × 106 4 × 10−18

Ecosphere 10−8 5 × 10−15% 2 × 107 1 2 × 10−34

a Openness, spatial scale and time scale are inverse to hierarchical scale.ariation on the molecular level.e number is the ratio of the two space scales in the exponent 2.

the molecular level is 100.

Ui

l

wl

deewcieiatdeiitmswf

Ftiqs

b Variation of the average contained component on the levels assuming a 100% vc Number of components contained in the next hiearchical level assuming that thd The value is on each level for an extensive variable assuming that the value on

.S. with many earthquakes taken place during a decade. The graphs a straight line in a double logarithmic plot, which entails:

og F = log a + b log M (2)

here as it can be seen from the graph (Fig. 3), that b = −0.8 andog a = 2.8 or a = 631.

Fig. 4 shows how the recovery time after a catastrophic eventecreases with the area size damaged. It means that the higher lev-ls in the hierarchy ecosystems, landscape, regions and the entirecosphere require longer time to recover, but it is not surprisinghen a larger area is damaged and the openness (expressed as

ircumference to area ratio) determining the dynamic is decreas-ng with the area. Fortunately, the frequency of bigger catastrophicvents is decreasing with the magnitude of the catastrophes. Fig. 5llustrates on the same graph the recovery time and the frequency,nd it can be seen they are opposite – they are so to say neu-ralizing each other. The probability for a catastrophe that wouldamage large areas – ecosystems, landscape, regions and the entirecosphere – is therefore decreasing with the hierarchical level. Ast was discussed above, the damping effect on the disturbance isncreasing with the hierarchical level. The hierarchical organiza-ion as a whole is therefore beneficial for the stability (survival or

aintenance) of the upper levels systems of the hierarchy. At theame time, have the lower levels (see Table 1) a very high dynamics,hich ensure a rapid renewal. The changes or the renewal is there-

ore a bottom-up effect, but due to the damping effect of the higher

Frequ enc y 1/y

0.01

100

10

1

0.1

4 5 6 1 2 3

Magn itude of earth quake Rich ter scal e

ig. 3. The frequency, F (events per year) is plotted versus the earthquake magni-ude, M, for an earthquake intensive area in U.S. for a decade. Notice that the plots double logarithmic, as the Richter scale is expressing the magnitude of an earth-uake in a logarithmic scale. As the graph is a straight line in log-log diagram, thehown relationship must be a power function: log F = log a + b log M.

Fig. 4. The approximate time for full recovery and the approximate frequency ofcatastrophic events are indicated versus the area damaged by the catastrophic event.

levels and the low frequency of disturbances, the survival of thehigher levels that control by feedbacks the lower levels is ensured.In other words the upper levels imposes constraints on the lowerlevels that will require an adequate and creative response fromthem in a dialectic manner (Nielsen, 2009). In this manner upperlevels induce a new agency at the lower levels (Ulanowicz, 2009).

The hierarchical organization is at the same time important for therenewal and development of nature and for the conservations ofparticularly the larger systems in nature.

Number of species

0 10 20 25

Dro

ug

h r

esis

tan

ce

-0.5

-1.0

-1.5

Fig. 5. The results of the experiments in grassland by Tilman and Downing (1994).The higher the number of species the higher the drought buffer capacity, althoughthe gain per additional plant species decreasing with the number of species.

5 ologic

e2o1et1Ipdrt

4h

1arcgavatewopafaldsaoi

eeeclessthboetIhpflttwsI

2 S.E. Jørgensen, S.N. Nielsen / Ec

Major catastrophes that have damaged significantly almost thentire ecosphere are know to have happened about 65 million,00 million and 251 million years ago. The frequency of a totalr almost total destruction of the ecosphere is about one time per00 million years. A very high percentage of all the species werextinguished by the three mentioned major catastrophes, and itook very long time – according to the fossils records more than0,000 years before the ecosphere was at least partially recovered.

t agrees with the trends in Fig. 4. The last major meteor strike hap-ened in Siberia in 1911, where more than 10,000 km2 forests wereestroyed. The landscape is today not yet fully, but only partiallyecovered. A meteor strike of this magnitude happens less than oneime per century.

. Selection of indicators that can account for theierarchical properties

Let us consider a landscape that has a size of00 km × 100 km = 10,000 km2. If it consists of only one ecosystem,n agricultural wheat field, the landscape is very vulnerable toandom disturbances as for instance the weather and pests. If itonsists let us say of 100 different ecosystems: wetlands, forests,rasslands, agricultural fields preferably of different types, lakesnd ponds, river and streams, the landscape would be much lessulnerable to drought or pest attack. It is probably accepted byll ecologists, but it is also understandable from the hierarchyheory, because the disturbances are very different in the differentcosystems and the disturbances will be damped in accordanceith (the number of ecosystems)0.5 or by a factor ten times in

ur illustration. In addition, the different ecosystems have alsoroperties that are able to cope with disturbances. A forest willttract precipitation and wetlands will absorb flooding. The dif-erent types of ecosystems offer different services and when therere many different ecosystems the spectrum of services by theandscape will of course be larger. The advantage of having 100ifferent ecosystems in our landscape case is therefore probablyignificantly underestimated by the use of a factor 10. It is therefores a consequence of this discussion proposed to apply the numberf different ecosystems in landscapes as an important ecologicalndicator for landscapes.

An ecosystem which has only few populations of species is alsovidently more vulnerable than an ecosystem that has many differ-nt species represented in many populations. If there is a randomlynvironmental disturbance for instance in the weather or climaticonditions, the species will have different reactions and that willevel off – damp – the overall disturbances of the ecosystems. Thexpected disturbance for the ecosystem with for instance 10,000pecies will be 50 times less than the ecosystem with only fourpecies, according to our rule of thumbs presented above. It meanshat the ecosystem with the highest number of species will with aigher propensity to maintain its functions when random distur-ances are causing impacts on the ecosystem. The different speciesffer furthermore a higher variety of functions (services) to thecosystem and the available services that are offered by an ecosys-em are therefore higher when the ecosystem has more species.n this context it should also be mentioned that different speciesave different sensitivity to different impacts. Copper is toxic tolants but at different levels and at different concentrations for dif-erent species. An ecosystem with many species will therefore beess sensitive to a moderate copper contamination because some ofhe plants will probably survive – or expressed differently, the pho-

osynthesis will with higher likelihood be retained in an ecosystemith many species, because there is higher probability with many

pecies that some of the species can cope with the contamination.t is consistent with the results obtained by Zavaleta et al. (2010)

al Indicators 28 (2013) 48–53

by an empirical model: higher species richness is required to pro-vide multiple ecosystem functions. To summarize: it is beneficialfor ecosystems to encompass as many different species as possi-ble, because this feature damps the random disturbances, give awider spectrum of services (consistent with Kunin and Lawton,1996) and a wider spectrum of different resistances to impactsfor instance to toxic substances. Species richness–the number ofspecies or Shannon’s index of species diversity – is therefore animportant ecological indicator for ecosystems. Generally, biodiver-sities on all the hierarchical levels are therefore to be consideredimportant indicators. A high biodiversity on all the hierarchical lev-els will contribute to a higher damping effect, to a wider spectrumof services, to a wider spectrum of resistances (buffer capacities)and a reduced vulnerability. See also Ehrlich and Ehrlich (1981,1992), where it is expressed as follows: all species make a contribu-tion to ecosystem processes, so that functioning of these processesdeclines progressively as species are lost.

It has shown experimentally that more diverse communitieshave more stable ecosystem functioning and have a consistentlyhigher level of functioning over time than less diverse ones. Morediverse plant communities have consistently higher productivity.Experiments over a period of seven years have given these clearresults; see Allen et al. (2011). The results are a clear experimentalsupport for the presented theoretical results of the damping/leveloff effect from one hierarchical level to the next due to the numberof components.

Tilman and Downing (1994) have presented another supportingexperiment: temperate grassland plots with more species have agreater resistance and buffer capacity (a smaller change in biomassbetween a drought year and a normal year) and have a faster recov-ery after drought; see Fig. 5.

As previously stated the findings by May (1981) are strictlycontradictory to the above mentioned importance of biodiversityand species richness. He uses mathematical analyses of patternsin multi-species communities and concludes, that increasing com-plexity in the sense of an increasing number of species andincreasing connectance or increasing average interaction strengthworks against dynamical stability. He shows also that mutualismtends to be a dynamically fragile relationship, which is in contrastto the 13 cardinal hypotheses about the properties of ecologicalnetworks (Patten, 1991; Jørgensen et al., 2007; Jørgensen, 2012).Furthermore, he mentioned ecosystems–for instance the marchdominated by Spartina, that are robustly enduring natural commu-nities, whereas complex K-assemblages maybe unable to recoveror that unprecedented perturbations imposed on ecosystems aremore traumatic for complex ecosystems than for simple ones. May’smathematical analysis is of course correct from a mathematicalpoint of view, but the systems that he is using are far too simpleto allow conclusions about the role of complexity or diversity forsuch complex systems as ecosystems. Even a monoculture is verycomplex when all the biological processes on all the hierarchicallevels are considered. The approach used in Sections 2 and 3 arevery different because it presumes a high number of componentsfor a hierarchical level and the mathematical stability is not used ascriteria, but the magnitude of the random disturbances in the var-ious hierarchical levels as function of the number of components.The examples mentioned above and taken from May (1981) aredirect observable, but it is not possible to conclude that a simplesystem is very stable and further that the diversity does not playa role or could even increase the stability. There are two compo-nents in these examples about the reaction of systems to impacts:the system and the impact. It means that a complex system can of

course be disturbed, damaged or even collapse, if the “right” impactis operating, but it does not imply that a complex system would bemore stable by decreasing the complexity or less stable by increas-ing the complexity. In this context, it should be emphasized that

ologic

twbFscPsep

5

tlovmtttttdtthnaottsb

R

A

S.E. Jørgensen, S.N. Nielsen / Ec

he results presented above show that the damping effect increasesith the number of components in a considered hierarchical level,

ut damping does not imply complete elimination of disturbances.urthermore, it should be emphasized that the increased diver-ity gives a probability for a wider spectrum of services and bufferapacities (resistances), but probability does not mean certainty.robabilities in ecology should preferably be understood as propen-ities sensu (Popper, 1990) as stressed by Ulanowicz (1986). So, thexamples mentioned by May (1981) can still be explained by theresented results in Sections 2 and 3.

. Summary and conclusions

It has been shown how it is possible to quantify openness andhereby understand the dynamics on the various levels of the eco-ogical hierarchy. The ecological hierarchy provides all the levelsf the hierarchy including the ecosystems and landscapes withery important properties. The higher levels can benefit from theore dynamic lower levels and at the same time provide impor-

ant regulations of the lower levels. The lower levels may due toheir dynamic change more radically by random disturbances, buthe hierarchical system is in most cases able to damp or level outhe changes in the lower level on the higher levels, which makeshe higher levels less sensitive to environmental disturbances. Theamping effect or level off effect is determined by the square root ofhe number in level n + 1 of the components from level n. It implieshat the diversity (number of components) in each level of theierarchy is an important indicator for the level, for instance theumber of ecosystems in a landscape and the number of species inn ecosystem. As the diversity increases, the spectrum of servicesffered by the various levels increases, too. Moreover, the spec-rum of resistances increases with the diversity. It seems thereforeo be crucial to use as indicators the number of ecosystems for land-capes and the number of species for ecosystems and generally theiodiversity for all the hierarchical levels.

eferences

llen, E., et al., 2011. More diverse plant communities have a higher functioning overtime due to turnover in complementary dominant species. Proc. Natl. Acad. Sci.U.S.A. 108, 17034–17039.

al Indicators 28 (2013) 48–53 53

Allen, T.F.H., Starr, T.B., 1982. Hierarchy, Perspectives for Ecological Complexity. TheUniversity of Chicago Press, Chicago and London, 310 pp.

Bak, P., 1996. How Nature Works. Springer Verlag, New York, 212 pp.Ehrlich, P.R., Ehrlich, A.H., 1981. Extinction: the causes and consequences of the

disappearance of species. Random House, New York.Ehrlich, P.R., Ehrlich, A.H., 1992. The value of biodiversity. Ambio 21,

219–226.Jørgensen, S.E., Fath, B., Bastiononi, S., Marques, J.C., Mueller, F., Nielsen, S.N., Patten,

B.C., Tiezzi, E., Ulanovicz, 2007. A New Ecology. Elsevier, Amsterdam, Oxford,276 pp.

Jørgensen, S.E., 2012. Introduction to Systems Ecology. CRC, Boca Raton/NewYork/London, 316 pp.

Kay, J.J., 1984. Self Organization in Living Systems [Ph.D. Thesis]. Systems DesignEngineering, University of Waterloo, Ontario, Canada.

Kunin, W.E., Lawton, J.H., 1996. Does biodiversity matter? In: Gaston, K.J. (Ed.), Bio-diversity: a Biology of Numbers and Difference. Blackwell, Oxford, pp. 283–308.

May, R.M., 1973. Stability and Complexity in Model Ecosystems. Princeton UniversityPress, Princeton, NJ, 265 pp.

May, R.M., 1981. Theoretical Ecology, second ed. Blackwell Scientific Publications,Oxford, London, pp. 197–227.

Nielsen, S.N., 2000a. Information in ecosystems. In: Jørgensen, S.E., Müller, F. (Eds.),Handbook of Ecosystem Theories and Management. Lewis Publishers, BocaRaton, FL.

Nielsen, S.N., 2000b. Thermodynamics of an ecosystem interpreted as a hierarchy ofembedded systems. Ecol. Model. 135, 279–289.

Nielsen, S.N., 2009. Thermodynamic constraints of life as downward causation inecosystems. Cybern. Hum. Knowing 16 (1), 27–49.

Nielsen, S.N., Müller, F., 2000. Emergent properties of ecosystems. In: Jørgensen,S.E., Müller, F. (Eds.), Handbook of Ecosystem Theories and Management. LewisPublishers, Boca Raton, FL, pp. 195–216.

O’Neill, R.V., DeAngelis, D.L., Waide, J.B., Allen, T.F.H., 1986. A hierarchical Conceptof Ecosystems. Princeton University Press, 286 pp.

Patten, B.C., 1991. Network ecology: indirect determination of the life–environmentrelationship in ecosystems. In: Higashi, M., Burns, T.P. (Eds.), Theoretical Ecosys-tem Ecology: The Network Perspective. London, Cambridge University Press, pp.288–351.

Patten, B.C., 1992. Energy, emergy and environs. Ecol. Model. 62, 29–69.Peters, R.H., 1983. The Ecological Implications of Body Size. Cambridge University

Press, Cambridge, 329 pp.Popper, K.R., 1990. A World of Propensities. Thoemmes, Bristol, 51 pp.Simon, H.A., 1973. The organization of complex systems. In: Pattee, H.H. (Ed.),

Hierarchy Theory—the Challenge of Complex Systems. Braziller, New York,pp. 1–27.

Tilman, D., Downing, J.A., 1994. Biodiversity and stability in grasslands. Nature 367,363–365.

Ulanowicz, R.E., 1986. Growth & Development: Ecosystems Phenomenology.Springer-Verlag, NY, 203 pp.

Ulanowicz, R.E., 2009. A Third Window. Natural Life Beyond Newton and Darwin.Templeton Foundation Press, West Conshohocken, PA, 196 pp.

Zavaleta, E., et al., 2010. Sustaining multiple ecosystem functions in grasslandcommunities requires higher biodiversity. Proc. Natl. Acad. Sci. U.S.A. 107,1443–1446.