Ecological Indicators 11 (2011) 97102

Contents lists available at ScienceDirect

Ecological Indicators
j our nal homepage: www. el sevi er . com/ l ocat e/ ecol i nd
Overview: Spatial information and indicators for sustainable management of
natural resources
Bertram Ostendorf

School of Earth and Environmental Sciences, The University of Adelaide, Adelaide, SA 5005, Australia
a r t i c l e i n f o
Article history:
Received 11 October 2010
Accepted 18 October 2010
Keywords:
Space-time
Decision support
DSS
GIS
Remote sensing
a b s t r a c t
Natural resource management (NRM) is becoming increasingly important at all scales, local, regional,
national andglobal, because of anincreasing humanpopulationandincreasing per capita use of resources
and space. Conicts are intensifying between different interest groups. Production and conservation
aspects are particularly debated because conservation often conicts with economic and social sustain-
ability. There is public demand for objective decision based NRM but limitations are all pervasive due to
the spatial and temporal complexity and interdisciplinary nature.
This special issue explores the use of spatial data and models to overcome some limitations of NRM
decision making. The papers in this issue show modern approaches of natural resources management
with a particular focus on spatial data collection, analysis and the development of spatial indicators. This
issue presents a balanced mix of reviewand research papers that give examples of howto nd or improve
the spatial information base for evidence-based decision making.
This overviewmakes the argument that understanding complex spatial pattern and processes, and the
development of spatial indicators, is an essential aspect of evidence-based NRM. If spatial and temporal
patterns are complex, ecological evidence fromeld data or experiments may have limited value for NRM
and observational study designs become more appropriate for understanding complex spatial pattern
and processes. Data quality should be documented as a combination of accuracy and spatio-temporal
representativeness in order to be useful in the NRM decision process.
2010 Elsevier Ltd. All rights reserved.
1. Introduction
Natural resource management (NRM) relates to the human
impact on the natural environment, the productivity of land and
water bodies and its impact on ecosystem services and qualities
such as water allocation, soil loss, biodiversity but also indirectly
with health issues as related to pollution, re, or dust storms. NRM
refers to maintaining quality of life and ethical values related to
sustainable management. It inuences ecosystems, landscapes and
becauseof off-siteimpacts andspatial interactions alsourbanareas.
With human population and resources use per capita on the rise,
goodmanagement is becoming increasingly important at all scales:
local, regional, national, and especially global. Debates amongst
different interest groups (e.g., producers and conservationists) of
management issues are intensifying as resources become limited
and because conservation of natural resources often conicts with
economic and social sustainability.
The increasing importance of NRM is paralleled with increas-
ing complexity. NRM issues are increasingly difcult to address

Tel.: +61 8 83037317; Fax: +61 8 83037617.

E-mail address: Bertram.Ostendorf@adelaide.edu.au
because historic inuences carry through into the future in an
ever increasing intensity andcomplexity. Past humanactions inu-
ence the ecosystem states and require continuing adaptation of
management (Argent, 2004). Management relies on continuously
improved and updated information, but is also highly dependent
on the dynamic nature of environmental conditions, most impor-
tantly; climate change, water allocation, soil loss and biodiversity
loss, with substantial interactions between these environmental
issues.
Human impact on the environment has been increasing in the
past and will do so into the future. The human footprint is becom-
ing larger because of population and standard of living increases,
whist natural resources are limited because of space and thermo-
dynamic constraints. The tragedy of the commons remains all
pervasive. Mismanagement potentially benets individual land-
holders in the short termwhile it adversely affects society at broad
scales and in the long term. But it also needs to be noted that
NRM actions are occurring at a property levels (often a very ne
spatial scale), whereas benets can only be felt if actions are coor-
dinated at much broader spatial and temporal scales. The recent
debate about local vs. broad scale governance related to biodiver-
sity conservation (Noss, 2010) indicates that resolutions require
further discussioninto the future. The inherent complexity of inter-
1470-160X/$ see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ecolind.2010.10.003
98 B. Ostendorf / Ecological Indicators 11 (2011) 97102
actions betweensocio-cultural, economical andbiophysical system
components makes management based on facts very difcult. Nei-
ther the future trajectory of the ecosystems nor outcomes, benets
and costs efciency of management actions may be predictable in
the future because the state of the system approaches levels that
we have not experienced before and the scales of inuence are
unprecedented in human history.
NRM recognises that many conicts arise because space on the
earth is limiting; NRMtherefore also is spatial management. There
is much evidence that NRM efciency is increased if spatial dif-
ferences are considered (e.g., Margules and Pressey, 2000, Naidoo
et al., 2006). The realism of management models is increasing
rapidly with the improvement of spatial data sources, models and
computational power. To be useful for NRM, realismand relevance
need to be high. The rich sources of some information layers at
high spatial and temporal resolutions may provide this ingredient.
Environmental decision support systems make use of increasingly
complex and integrative model structures, user friendliness and
are increasingly used in stakeholder meetings, which provide the
basis for objective discussions and negotiations (Matthies et al.,
2007).
This special issue contains a selected set of papers that were
presented in the session Spatial Data for NRM at the Ecosummit
2007. Papers range from terrestrial to marine and cover key areas
of NRM (devising of management units, biophysical and biologi-
cal condition assessment, generation of productivity and economic
information layers and development of models for evaluation of
management options). The articles in this issue show modern
approaches to natural resources management with a particular
focus on spatial information collection, analysis and the develop-
ment of spatial indicators. Both review and research papers show
studies that improve spatial informationfor the NRMdecisionmak-
ing process. The complexity of many NRMissues demands complex
linkages of models and data for assessment of realistic future sce-
narios and evaluation of alternative management options.
2. What factors limit objective NRM decision making?
The strength of a chain of decisions depend on the weakest link
of the available information. Too often critical spatial information
layers are difcult to come by at the appropriate spatial resolution
and extent, hence reducing the overall strength of the combined
evidence. Many processes (e.g., carbon, water, nutrients and pollu-
tants transport) or patterns that are either directly of management
relevance or control critical processes are difcult or impossible to
measure at the most relevant spatial or temporal scales. Decisions
rely on more or less substantiated assumptions and are limited
by our lack of understanding of the temporal dynamics of spatial
pattern and their trajectories into the future.
3. Causes of spatial patterns and processes
Spatial complexity is apparent at all scales and humans play a
key role shaping the land (e.g., Tasser et al., 2008). Simple visu-
alisation of surface reectance from satellite imagery shows the
most obvious causes of pattern (Fig. 1). Broad scale spatial pattern
that are visible at scales from1:100 millionto 1:10 millionare most
strongly inuencedby geology, topography andglobal climatic pat-
tern as inuenced by latitude and geography of continents. Human
inuence, even of such densely populated metropolitan areas as
Beijing, China is not yet strongly evident. At scales of 1:1 million,
humanactions are clearlyvisible inhighlypopulatedareas; but also
Fig. 1. Earth surface pattern at different scales (source Google Maps). Whereas at continental and global scales pattern are mostly natural, at 1:1 million the inuence of
humans on shaping land surfaces are one of the strongest factors.
B. Ostendorf / Ecological Indicators 11 (2011) 97102 99
anthropogenic agricultural cropping and forestry pattern become
visible. At a 1:500,000 scale, consequences of human action are
dominant.
Spatial differences in conditions may require specic man-
agement. One important basis for all decision processes is the
delineation of spatial units that are relatively homogeneous and
can be managed as a whole. In addition to the practical purpose for
zonal management, identication of management units may also
help to optimise locations for eld data collection or monitoring
to increase spatial representativeness. It may also allow indicator
development that would allow spatial surrogates or proxies to be
assessedregularly, at a highspatial resolutionandat a broadspatial
extent.
The most obvious choice is to apply classicationmethods using
data that describe the heterogeneity of the region of interest. This
problem is addressed by two papers in this issue.
Bierman et al. (2011) review the applicability of classication
methods for remote sensing imagery, exemplied for marine water
quality. The authors conclude that there is a wide variety of meth-
ods that can be readily applied and with the wide availability of
vast satellite image archives time series analysis of spatial data in
classication approaches is facilitated. Pesch et al. (2011) present a
national approachto regionalisationfor bothmarine andterrestrial
environments. They suggest that the use of the decision tree algo-
rithmClassication and Regression Trees (CART) is equally feasible
for the broad-scale classication of marine and terrestrial condi-
tions. The usefulness of the classicationresults is demonstratedby
their ability to stratify temporal information of pollutant monitor-
ing and temperature dynamics in terrestrial and marine examples,
respectively. Zones describe spatial combinations of biotic and abi-
otic characteristics and are useful for management, analysis, and
optimal location of monitoring sites; spatial modelling that is done
before data collection will likely improve cost efciency of man-
agement.
4. The source of evidence for decision making
Decision making in natural resources management depends on
the appropriateness of the available data. Lack of informationat the
appropriate spatial resolution and/or extent, reduces the overall
strength of the combined evidence (e.g., Parrott, 2010; Proulx and
Parrott, 2009).
With increasing spatial complexity, the number of factors and
factor levels that need to be examined increases rapidly. Num-
bers may become unmanageable, too costly or simply impossible
for logistic reasons. There is substantial debate about how to
provide the evidence that is necessary for evidence driven manage-
ment. Havens and Aumen (2000) suggests that hypothesis-driven
research needs to underpin management as condence in the
decisions can only be gained by clear, structured and testable
approaches. Information from controlled experiments provides a
solid, testable understanding of systems, but with increasing spa-
tial complexity, such information only provides limited return for
the NRMdecision making process showing the limits of traditional,
reductionist scientic inquiry. If study areas are spatially complex,
the value of experimental research for decision making decreases
as experiments cannot represent actual pattern at the extent of the
management scale. The risk of erroneously generalising from spa-
tially non-representative samples (e.g., by making the assumption
of spatial stationarity) needs to be balanced with the risk of iden-
tifying relationships that may not be causative. The value in the
decision making process of indicators, even if they contain large
condence bands, increases along a gradient of increasing spatial
complexity; here observational study designs become more appro-
priate than manipulative experiments for understanding complex
Fig. 2. With increasing spatial complexity evidence generated from manipulative,
experimental study designs is reduced, whereas evidence generated from spatial
indicators, even if correlative and not causative, increases.
spatial pattern and processes and studies that relate patterns to
processes become increasingly valuable (Fig. 2).
Soils are an extreme example of spatial heterogeneity. Soil con-
ditions affect the entire range of ecological processes leading to
differences in productivity and biodiversity, the soil matrix medi-
ates processes that are the basis for key ecosystem properties
related to water and nutrients. The soil matrix determines a wide
range of biophysical properties for a large number of critical goods
and services of ecosystems. But soils are inherently variable in
space. Even the most detailed maps smooth over substantial het-
erogeneity. Problems of samplingscaleareeminent. Samplinggrain
(size of sample that is taken) needs to be large as variability is high
at very ne spatial scales, and sampling frequency needs to be high
because variability is high along climatic, vegetation, topographic
gradients, geological pattern (Jenny, 1941).
It is clear that improved spatial representations of soils would
benet NRM. Summers et al. (2011) explore the potential of hyper-
spectral remote sensing to assess soil conditions at higher spatial
and attribute detail. Their results from laboratory studies show
signicant relationships between spectral and physicochemical
properties. This has the potential to reduce the time for soil core
analysis, hence increase sample numbers. In addition, results pro-
vide supporting evidence for the use of hyperspectral remote
sensing to assess detailed spatial patterns of soil surface properties.
Spatial data of ecological conditions are equally difcult to
assess inthe eld. For one, they are linkedto soil conditions andsoil
variability. In addition, ecosystems show strong internal dynamics
that affect plant community patterns in ways even more complex
than soils. A high level of unexplainable variability (noise) exists
because it is in principle impossible to trace the history of plant
development and/or interactions that arise from stochastic events
such as death, disturbance or germination. Additionally, historic
differences in resource availability may show a lagged response.
Consequently, an even higher number of samples or larger sam-
pling grain might be necessary to identify pattern, as the signal to
noise ratio is reduced, imposing very strong logistic constraints on
eld sampling.
Clarke et al. (2011) evaluate the representativeness of vegeta-
tion sampling in arid grazing regions. Using comprehensive data
sets from two different independent government agencies they
found that biodiversity estimates strongly depend onthe collection
effort. Rarefaction could not remove the bias even if analyses were
restricted to perennial plants. The paper of Clarke et al. (2011) is an
example that shows that even the most comprehensive and costly
eld surveys still contain issues that may need to be corrected.
Structural ecosystem properties are less inuenced by popula-
tion dynamics than species composition because structure is more
dependent on external, climatic drivers that vary systematically
100 B. Ostendorf / Ecological Indicators 11 (2011) 97102
along topographic and broad scale geographic gradients. Chen et al.
(2011) nd that aboveground forest biomass carbon can be accu-
rately estimated by using satellite imagery as a source of detailed
spatial pattern. Combining comprehensive spatial ground-based
and modelled information layers (including detailed forest ser-
vice data and spatial re pattern models), the authors were able
to develop strong empirical regression models of aboveground for-
est biomass carbon. The study provides the basis for predictions
of important management variables from readily available spa-
tial information layers. More importantly, these regression models
allow assessment of the dynamics of very detailed management
information layers if older imagery is used in the models. With
increasing length of time-series imagery, it is becoming possible
to assess long-term changes, and with this information also make
inferences about trajectories into the near future with high con-
dence.
5. Temporal and spatial non-stationarity
Field data or experimental evidence is very difcult to come
by for NRM because of the temporal variability in spatial pat-
terns of all ecological conditions and biophysical drivers causing
change. This is particularly evident for arid regions, where rain-
fall and re are two highly stochastic drivers with a long memory
effect on vegetation pattern, primary production and all dependent
ecosystem processes. Management and monitoring need to con-
sider this variability. Furthermore, rangeland management (e.g.,
devising conservation areas, dening maximumstocking rates and
monitoring) is done over a large spatial extent. Field data are inher-
ently difcult to obtain (see also Clarke et al., 2011; Feng and Zhao,
2011; Jafari et al., 2008; Turner et al., 2008).
Turner et al. (2011) examine spatial and temporal variability of
reinaridAustralia. Their results showmarkedregional differences
of joint inuences of rainfall of differing antecedent periods, veg-
etation and soils. The authors have developed a conceptual model
which helps to separate potential factors and allows separate anal-
ysis of causal inuences. Their results show that rainfall history
one of the most variable factors, is most important. Understanding
temporal dynamics and scales of spatial pattern can support man-
agement by stratifying the study area into meaningful domains of
similar response. Uncertainties in the models also help to identify
data gaps and improvements of long-lead predictive models.
Most articles in this issue deal in different ways with the prob-
lemof time-space variability. Fig. 3 further illustrates this problem
using a marine example of chlorophyll-a estimations in the water
column from MODIS satellite imagery. Even if we were able to
measure the spatial variability of chlorophyll concentrations along
transects with high accuracy, we would still be sampling condi-
tions of a chaotic system- a useless waste of moneyandeffort. More
importantly, a single time slice does not tell the complete story and
cannot be usedtodrawgeneral conclusions that wouldbe useful for
management such as identication of locations with a higher risk
of algae blooms. However, time series of satellite imagery reveal
the signal of time-space variability and can provide much more
useful information than a single static image or eld survey. On the
other hand, it is also evident that the temporal variability of a single
location strongly depends on chaotic spatial processes, which are
impossible to predict (similar to predicting the weather conditions
more than a few days into the future), but can be remotely sensed.
6. NRM decision making into the future - Using spatial
models in the NRM decision making process
Future conditions may be different from conditions that a sys-
tem has experienced in the past. This makes objective decision
support a very difcult task as decisions need to rely on models.
Models can be dened as abstractions from reality that are based
on the essential components of a problem. Model development is
Fig. 3. Satellite imagery of consecutive daily Chlorophyll-a pattern in the Spencer Gulf, South Australia. The dark smooth grey area is the land mask; cloud cover or missing
data are shown in white.
B. Ostendorf / Ecological Indicators 11 (2011) 97102 101
an iterative process that considers availability of data, understand-
ing of processes, and identication of temporal and spatial scales.
The modelling process further requires an identication of critical
model components as well as the development of new, or adaption
of existing modelling techniques.
Models in NRMare ultimately concerned with decision support
at the interface of human exploitation of resources and environ-
mental integrity. This necessitates integrated approaches where
economic and social aspects are combined with factors related
to biophysical conditions of the disputed regions. Assessment of
present, andprojections of future conditions as moderatedby man-
agement decisions, need to be considered in a spatial context.
Three papers in this issue consider the full suite of assessment
and modelling and illustrate the complexity of models and data
that need to be combined if realistic future scenarios of manage-
ment options are to be assessed. Burkhard et al. (2011) show an
example of marine and coastal management models. The prob-
lemdemands ecological integrity indicators that combine complex
information at different spatial and temporal scales. An integration
of a number of key processes was successfully accomplished and
spatial biological observations were linked with simulation models
of the marine food chain, biochemical sediment and hydrodynamic
processes. The model linkage shows a complex ecosystemresponse
over time withpotential positive as well as negative impacts andan
overall resilient nature of ecosystemfunctions. Realistic results are
only possible through a spatial framework, where the spatial and
temporal heterogeneity of all key inuential factors and boundary
conditions can be presented in their actual complexity and inte-
grated. This makes it possible to specically suggest future work
and necessary enhancement of models and their data base.
Feng and Zhao (2011) have developed a model for grazing man-
agement in the steppes of northern China. The vast extent of the
management region demands modelling and monitoring based
on satellite imagery. The model complex integrates a prominent
biogeochemical point model (CENTURY) with satellite imagery
in order to manage overgrazing and permanent soil loss, and
to prevent dust storms. Management needs to consider complex
interactions between economic, social and ecological sustainabil-
ity, the effect on global atmospheric CO
2
concentration, and the
potential negative impact on and from global climate change.
A combination of static spatial maps of soil conditions that are
assumed homogeneous and satellite imagery derived indices was
usedtodrive the model. Most innovatively, aniterative process that
compares modelledandobserved(satelliteimagery) patternallows
separationof the inuence of climate, soils and grazing intensity on
land surface conditions and thus provides an indicator for spatial
differences in local land management.
Crossman et al. (2011) similarly suggest that it is impor-
tant to combine multiple spatial datasets and models to improve
management and highlight the need for interdisciplinary work.
Distribution and abundance of weeds impose substantial cost for
agricultural production and are a severe threat for native ora and
fauna. Assessment of weed threats is difcult due to interactions
of spatial conditions with climate change. Management needs to
consider a large number of factors and uncertainties that arise from
attribute, spatial, but also methodological uncertainties (addressed
by using different statistical models).
Ultimately, all models depend on evidence that is based on data
samples that were created at specic locations and specic time
intervals. Usefulness andaccuracy of model predictions depends on
the appropriateness of temporal and spatial scales at which infor-
mation has been gathered. Whilst it is the ultimate applied goal for
models to directly support management, spatial indicator devel-
opment may also consider development of those variables that are
essential components in such integrated models. Models also serve
as tools for the development of data layers that ll critical gaps in
the NRM decision making process. A large number of data layers is
either too difcult or impossible for legal or theoretical reasons to
obtain at appropriate detail and/or extent.
One of the most important data layers in NRM is economic via-
bility. This is addressed in the contributions by Bryan et al. (2011)
and Lyle and Ostendorf (2011). NRM actions are ultimately exe-
cuted by land managers, and as most land is managed primarily
for a sustained income, even in multifunctional landscapes (e.g.,
Willemen et al., 2010, Gottschalk et al., 2010) decisions are largely
driven by economic considerations. Changes in income that arise
from abandonment of current land-use practices must be at least
compensated by alternative income in order to be accepted. The
adoption of alternative management by land holders will depend
on the continued ability to make prot. Policy development would
strongly benet froman ability to anticipate propensity for change
of landmanagers. Bryanet al. (2011) demonstratetheneedtoassess
agricultural opportunity as a spatially andtemporally variable indi-
cator. Using statistical zones as spatial units, they showsubstantial
variability with the state of South Australia in income and hence
opportunity loss that would need to be considered in environmen-
tal policy development. The ability to map the uncertainty adds a
further information layer supporting management decisions.
The contribution of Lyle and Ostendorf (2011) shows that it is
possible to generate economic informationlayers at ne detail over
broad spatial scales. Farm management increasingly uses detailed
yield maps for precision agriculture. Variability within elds is
often very high and targeted management might allow identi-
cation of sections within elds that either generate little income or
produce losses. These areas might be targeted for alternative land
uses, such as biodiversity or carbon storage plantations. Regional
management agencies would have an increased ability to antici-
pate landholder response to policies if they had information layers
that had a similar detail as the information used for on-ground
decision making by landholders. The authors have successfully
related land surface reectance to economic return from land for
cropping, hence allowing information about sub-eld variability
(withinmetres) ineconomic performance tobe producedat a state-
wide extent (millions of square kilometres). It is therefore possible
for regional managers to obtain information at the extent of their
entire management area with a spatial detail that otherwise only
landholders would have to manage their farms.
7. Summary
Increasing scarcity of natural resources (water, soil, vegetation)
coincides with an increasing demand from many interest groups
(e.g., urban expansion, industry, agriculture, recreation, conserva-
tion), underpinning the needs for an evidence-driven management
of our natural resources. This special issue demonstrates that NRM
requires spatial indicators of states and processes and that there is
a need for an increased awareness of scales and variability in order
to move from subjective to objective evidence-based NRM.
The quality of decision making in natural resources manage-
ment depends on the appropriateness of the available data, and
the appropriateness of theories that underpin management deci-
sions. Thestrengthof thedecisionoutcomedepends ontheweakest
link of all data layers, but too often critical spatial information lay-
ers are difcult to come by at the appropriate spatial resolution
and/or extent, hence reducing the overall strength of the com-
bined evidence. Not only do we need to deal with comprehensive
linkages betweennatural (ecological), economic andhuman(socio-
political-institutional) subsystems, but we are also working in a
situation that is highly dynamic in time and space.
Decisions should be made based on evidence, or in other words
on solid and objective information. But the reality is often differ-
102 B. Ostendorf / Ecological Indicators 11 (2011) 97102
Expenditure on data collecon
B
e
n
e

t

f
o
r

d
e
c
i
s
i
o
n

m
a
k
i
n
g
Misrepresentaon of spao - temporal paern
Fig. 4. Data collection is not benecial for the decision making process if spatio-
temporal patterns are misrepresented.
ent. Frequently, information is either detailed and accurate but not
widely applicable, or widely applicable, but not accurate enough
to give condence in its use. Articles in this issue support the
argument that there is need for an increased awareness of spatial
and temporal representativeness in the core scientic studies that
underpin the decision making process and that current advances
in theories, methods and data availability are substantial.
Tobeuseful inNRMdecisionmaking, studies needtobespatially
representative. Hard evidence from experimental studies can only
be collected at a limited number of factor combinations. This num-
ber is further reduced if experimental treatments are conducted
using eld trials with a higher realism than laboratory studies.
Experimental eld studies can only be limited to a few locations.
This information is only useful if it is collected at temporal or spa-
tial scales that reect the natural variability of patterns and the
variability over time. If spatial patterns are misunderstood or mis-
representedintheselectionof elddataor inthechoiceof sampling
grain, information may not be benecial for the decision making
process. In fact, it may be costly in that the condence that comes
with the increased effort may force us to falsely assume that the
available information provides a solid basis for decisions (Fig. 4).
Informationthat has beenobtained at inappropriate spatial or tem-
poral scales, will adversely affect decision outcomes.
Field data, even if highly accurate, may only truly represent
conditions at one point in time at a single location. It would be
pointless to use rainfall measurements in the arid regions at one
location and a single time period to assess long-term water avail-
ability, but vegetation surveys taken at one point in time are often
the only information available to characterise site conditions and
monitor change. Caution is necessary. If spatial and temporal rep-
etitions are too small, ecological eld data (even if locally highly
accurate and repeatable), may be of very little value for regional
assessment and rather constitute an expensive method to generate
random numbers. Data quality should be documented as a combi-
nation of accuracy and spatio-temporal representativeness. While
there are calls for better documentation of data accuracy (e.g., Boyd
and Foody, in press), representativeness of data or experiments has
received much less attention (see also Cushman and Huettmann,
2010). Papers in this issue support the argument that spatial repre-
sentativeness is of high importance if local solid experimental and
observational evidence is to be used in broad scale management.
It is encouraging in this respect to see a large number of jour-
nal articles that explicitly consider spatial issues in their research.
About 270,000 references are listed in the Scopus database that
use the terms ecological or ecology in the title, keywords or
abstract. The proportion of these papers with the term spatial has
quadrupled from 3% in the 1960s and 1970s to more than 12%
between 2000 and 2010. The overall proportion of papers in the
journal Ecological Indicators that include the term spatial has
been 20% with an increase to 25% during the last two years.
In order to obtain the best future outcomes, we need to assess
theeffect of management actions andcontrol theseaccordingly. But
in order to control, we need to understand local, current, regional
off-site and future consequences of policies and actions. Spatial
information is fundamentally important for the development of
indicators of conditions and as well as for developing decision pro-
cesses for sustainable NRM into the future.
References
Argent, R.M., 2004. An overview of model integration for environmental applica-
tions Components, frameworks and semantics. Environmental Modelling and
Software 19, 219234.
Bierman, P., Lewis, M., Ostendorf, B., Tanner, J., 2011. A review of methods for
analysing spatial and temporal patterns in coastal water quality. Ecological
Indicators 11, 106117.
Boyd, D.S., Foody, G.M., Anoverviewof recent remote sensingandGIS basedresearch
in ecological informatics. Ecological Informatics In Press, Corrected Proof.
Bryan, B.A., King, D., Ward, J.R., 2011. Modelling and mapping agricultural oppor-
tunity costs to guide landscape planning for natural resource management.
Ecological Indicators 11, 202211.
Burkhard, B., Opitz, S., Lenhart, H., Ahrendt, K., Garthe, S., Mendel, B., Windhorst,
W., 2011. Ecosystem based modeling and indication of ecological integrity in
the German North SeaCase study offshore wind parks. Ecological Indicators
11, 171177.
Chen, X., Liu, S., Zhu, Z., Vogelmann, J., Li, Z., Ohlen, D., 2011. Estimating aboveground
forest biomass carbon and re consumption in the U.S. Utah High Plateaus using
data from the Forest Inventory and Analysis Program Landsat, and Landre.
Ecological Indicators 11, 143151.
Clarke, K., Lewis, M., Ostendorf, B., 2011. Additive partitioning of rarefaction curves:
Removing the inuence of sampling on species-diversity in vegetation surveys.
Ecological Indicators 11, 135142.
Crossman, N.D., Bryan, B.A., Cooke, D.A., 2011. An invasive plant and climate change
threat index for weedrisk management: Integrating habitat distributionpattern
and dispersal process. Ecological Indicators 11, 186201.
Cushman, S.A. and Huettmann, F., (eds.) 2010. Spatial Complexity, Informatics, and
Wildlife Conservation, Springer Tokyo Berlin Heidelberg New York. 458pp.
Feng, X.M., Zhao, Y.S., 2011. Grazing intensity monitoring in Northern China steppe:
Integrating CENTURYmodel and MODIS data. Ecological Indicators 11, 178185.
Gottschalk, T.K., Dittrich, R., Diekotter, T., Sheridan, P., Wolters, V., Ekschmitt, K.,
2010. Modelling land-use sustainability using farmland birds as indicators. Eco-
logical Indicators 10, 1523.
Havens, K.E., Aumen, N.G., 2000. Hypothesis-driven experimental research is nec-
essary for natural resource management. Environmental Management 25, 17.
Jafari, R., Lewis, M.M., Ostendorf, B., 2008. Animage-baseddiversity index for assess-
ing land degradation in an and environment in South Australia. Journal of Arid
Environments 72, 12821293.
Jenny, H., 1941. Factors of Soil Formation. A System of Quantitative Pedology.
McGraw Hill Book Company, New York, NY, USA, pp. 281.
Lyle, G., Ostendorf, B., 2011. A high resolution broad scale spatial indicator of grain
growing protability for natural resource planning. Ecological Indicators 11,
212221.
Margules, C.R., Pressey, R.L., 2000. Systematic conservation planning. Nature 405,
243253.
Matthies, M., Giupponi, C., Ostendorf, B., 2007. Preface - Environmental decision
support systems: Current issues, methods and tools. Environmental Modelling
& Software 22, 123127.
Naidoo, R., Balmford, A., Ferraro, P.J., Polasky, S., Ricketts, T.H., Rouget, M., 2006.
Integrating economic costs into conservation planning. Trends in Ecology and
Evolution 21, 681687.
Noss, R.F., 2010. Local priorities can be too parochial for biodiversity. Nature 463,
4241424.
Parrott, L., 2010. Measuring ecological complexity. Ecological Indicators 10,
10691076.
Pesch, R., Schmidt, G., Schroeder, W., Weustermann, I., 2011. Application of CART
in ecological landscape mapping: Two case studies. Ecological Indicators 11,
118125.
Proulx, R., Parrott, L., 2009. Structural complexity in digital images as an ecological
indicator for monitoring forest dynamics across scale, space andtime. Ecological
Indicators 9, 12481256.
Summers, D., Lewis, M., Ostendorf, B., Chittleborough, D., 2011. Visible near-infrared
reectance spectroscopy as a predictive indicator of soil properties. Ecological
Indicators 11, 126134.
Tasser, E., Sternbach, E., Tappeiner, U., 2008. Biodiversity indicators for sustainability
monitoring at municipality level: An example of implementation in an alpine
region. Ecological Indicators 8, 204223.
Turner, D., Lewis, M., Ostendorf, B., 2011. Spatial indicators of re risk in the arid
and semi-arid zone of Australia. Ecological Indicators 11, 152170.
Turner, D., Ostendorf, B., Lewis, M., 2008. An introduction to patterns of re in arid
and semi-arid Australia, 1998-2004. Rangeland Journal 30, 95107.
Willemen, L., Hein, L., van Mensvoort, M.E.F., Verburg, P.H., 2010. Space for peo-
ple, plants, and livestock? Quantifying interactions among multiple landscape
functions in a Dutch rural region. Ecological Indicators 10, 6273.