Why is the Study Area Important in SDM?


In species distribution modelling (SDM), the study area defines the spatial extent over which your environmental predictors and species occurrence data will interact. A well-defined study area ensures:

  • Biologically realistic predictions

  • Proper ecological constraints

  • Avoidance of extrapolation into unsuitable or irrelevant areas


Choosing an inappropriate extent may introduce biases or overgeneralisation in model outputs (Elith & Leathwick, 2009; Zurell et al., 2020).

Study Area Options in EcoCommons


EcoCommons currently offers four primary methods to define your study area in the SDM workflow. These options cater to different ecological, methodological, and logistical needs.


1. Convex Hull


Description:

The convex hull is a minimum polygon enclosing all your species’ occurrence records, the smallest area containing all points without any internal angles.


Use case:

  • When occurrences are tightly clustered

  • When you want to constrain the model to only areas observed


⚠️ Limitation: May exclude ecologically relevant but unsampled areas.


Example: Species with limited sampling may have a convex hull that doesn’t reflect their full range.

2. Pre-defined Extents (e.g., NRM, IBRA, States)


Description:

Choose from administrative and ecological boundaries such as:




Region Type 

 

What it is / how defined 

Key features / what’s included 

Best Use in SDM / Ecological Modelling 

World Country Boundaries 

Political/administrative boundaries of countries globally. Typically, defined using ISO 3166 country codes, these boundaries are used in international reporting, geopolitical analysis, and global biodiversity assessments. 

Includes all UN-recognised countries and dependent territories. Often generalised to match scale (e.g. small island states may be exaggerated or simplified). Boundaries reflect administrative claims, not necessarily ecological or biogeographic regions. 

Best used in global or continental-scale studies, such as modelling invasive species spread, migratory bird distributions, or climate-driven range shifts across borders. Also used in reporting to international conventions (CBD, IPBES, etc). Less ecologically meaningful within-country. 

Local Government Areas (LGAs) 

 

Administrative boundaries of local councils / shires etc. Each state/territory divides its land into LGAs; these are political/administrative rather than ecological. 

Very detailed units, variable in size. Includes urban, rural, mixed-use. Political responsibilities (local infrastructure, land use, some environmental regulation). 

Good for when policy, local stakeholders, or management plans are defined at LGA level; for finer scale work, or when occurrence data is dense and you want fine resolution. But less useful if you want ecologically coherent units. 

National Resource Management (NRM) Regions 

 

These are administrative/ecological regions used by Australia for planning, managing land, water, biodiversity etc. There are 54 regional NRM organisations. Boundaries are based on catchments, bioregions, sometimes community/geographic features. 

Covers whole of Australia (land, coastal, islands) where NRM activities take place. These regions are used for natural resource funding, policy, planning. Sometimes overlapping with other classification systems (e.g. IBRA) but oriented around management and governance. 

Very useful when aligning modelling with policy, funding, conservation or restoration planning. If stakeholders are NRM agencies, or if you need to produce outputs in the units they use, this is a good choice. 

Australian States and Territories 

 

The political divisions of Australia: e.g. New South Wales, Queensland, Western Australia, Northern Territory, etc. 

Very broad extent; may include many ecosystems, environmental gradients. Very coarse for ecology, but aligned with many legal/policy jurisdictions. 

Useful as baseline extent if policy is state‑based, or if predictors are available/state level; but ecological variability within states is high. Not great for species tied to narrow ecological niches. 

IBRA Bioregions (e.g. “IBRA 7 Regions”) 

 

Interim Biogeographic Regionalisation for Australia (IBRA) is a landscape classification of Australia into regions & subregions that share similar geology, climate, landform, vegetation, and species assemblages. 

Contains 89 large-scale bioregions + 419 subregions in Version 7. Covers terrestrial Australia (excluding Antarctica). Examples: Wet Tropics, Australian Alps, Nullarbor Plain. 

Useful when you want your model to be ecologically meaningful in terms of terrestrial habitats; good for assessing representation, conservation planning, or when you want to limit the study area to known vegetation/geomorphology-based boundaries rather than administrative limits. 

River Regions/ Drainage Divisions Level 1 & 2 

Based on hydrology: major drainage divisions (large-scale), subdivided into river regions / basins. These are geographic units defined by water flow, catchments, topography. Includes “Drainage Divisions Level 1/Level 2” in many spatial datasets. 

For example: Australia’s River Basins 1997 divides into 12 (or more) major drainage divisions, then water regions, then many river basins. River Regions are often finer-grained units derived from drainage basins. 

If your species’ distribution is strongly tied to water (aquatic, riparian, wetland species), or where hydrology is a key limiting factor (flow, flood regimes etc.), using drainage/rivers-based extents makes sense. Also helps align with water‑resource management planning. 

Marine Ecoregions of the World 

Global marine ecoregion scheme (coastal & offshore), such as those used by WWF, etc. These classify marine zones globally by ecological frameworks. 

Covers entire world’s marine zones; includes categories like coral reefs, temperate shelves etc. Useful for global or coarse-scale marine work. 

When you want a broader marine context (beyond Australia or comparing across countries), or when projecting climate change effects; or for marine-only species with broad ranges. 

IMCRA 4 Bioregions (Provincial & Meso‑scale) 

 

Marine / coastal regionalisation: the marine equivalent of bioregional classification, based on biological + physical + oceanographic data (seafloor geomorphology, species distribution, sediments etc.). Two scales often used: “provincial” (broad) and “meso‑scale” (finer). Version 4.0 is the current major Australian marine scheme. 

Provincial bioregions: fewer, larger marine units; meso-scale: more many smaller units, especially in nearshore zones. Covers coastal waters to the edge of the Exclusive Economic Zone (EEZ), excluding remote territories sometimes. 

Use when modelling marine / coastal species, or interactions between terrestrial-marine (e.g. riparian, estuarine species). Helpful if environmental predictors include marine variables (sea temperature, currents, salinity etc.). Also useful for marine protected area planning. 


Use case:

  • Aligning with conservation planning units (e.g., NRM targets)

  • Reporting to government or regional bodies

  • Focusing models within known ecological contexts


Tip: Use IBRA or NRM zones for species tied to particular vegetation communities or habitat types.


Example: A researcher focusing on Stuttering Frog may choose 'Sydney Basin' IBRA 7 region to reflect its core habitat.

3. Bounding Box of Predictor Data


 Description:

Automatically defines the study area using the spatial extent of your environmental predictor layers.


 Use case:

  • Ensures spatial consistency between predictors and model domain

  • Useful when your climate/vegetation/fire layers are already clipped to a region


⚠️ Limitation: May be too broad if predictors include large buffers or national-scale datasets.


Example: If your bioclim data is clipped to Southeast QLD, the bounding box will follow this.

4. Draw Extent on Map (Manual Selection)


Description:

Use a visual map interface to draw a custom polygon over your desired modelling region.


Use case:

  • Custom conservation projects (e.g., park boundaries)

  • Areas of recent field surveys

  • Testing SDMs across gradients (e.g., altitudinal zones)


Pro Tip: Combine this with local knowledge of habitat boundaries.

5. Upload Shapefile (Coming Soon)


Important Considerations / Trade‑offs


When choosing among these regions as your “study area” in a modelling experiment, here are some scientific trade‑offs to consider (these are useful to include in your article as “how to choose” or “what to watch out for”):

  • Ecological coherence vs administrative convenience
    Ecologically defined regions (IBRA, IMCRA, drainage divisions) tend to better align with environmental gradients, species distributions, etc. Administrative regions (states, LGAs, NRM) may be more convenient for data access, governance, reporting—but they may mix very different ecosystems.

  • Scale & resolution
    The finer the region, the more precise you might be, but also potentially more data noise and edge effects. For example, using LGAs or small river basins could be great if occurrence data is dense and environmental predictors exist at matching resolution. But if data is sparse, using coarse regions could reduce overfitting.

  • Predictor coverage
    Whatever region you pick, ensure your environmental predictor datasets cover that extent. If you pick a region where some predictors are missing or coarsely resolved, your model may extrapolate poorly.

  • Edge / boundary effects
    Boundaries (especially in hydrology or marine regions) may cut through key features. For example, a basin boundary may exclude upstream habitat that matters; marine provincial boundaries might not align with species’ larval dispersal zones.

  • Data availability & stakeholder alignment
    If conservation or policy partners expect results by NRM or by State, aligning region now means your outputs are easily comparable. Also, data (occurrence or environmental) may be easier to source when aligned with known region units.


Citation:


Department of Climate Change, Energy, the Environment and Water. (2023). Interim Biogeographic Regionalisation for Australia (IBRA), Version 7. https://www.dcceew.gov.au/environment/land/nrs/science/ibra


Elith, J., & Leathwick, J. R. (2009). Species distribution models: Ecological explanation and prediction across space and time. Annual Review of Ecology, Evolution, and Systematics, 40, 677–697. https://doi.org/10.1146/annurev.ecolsys.110308.120159


Araújo, M. B., & Peterson, A. T. (2012). Uses and misuses of bioclimatic envelope modeling. Ecology, 93(7), 1527–1539. https://doi.org/10.1890/11-1930.1


Zurell, D., Franklin, J., König, C., Bouchet, P. J., Dormann, C. F., Elith, J., ... & Merow, C. (2020). A standard protocol for reporting species distribution models. Ecography, 43(9), 1261–1277. https://doi.org/10.1111/ecog.04960


Natural Resource Management Regions Australia. (2021). What is Natural Resource Management? https://nrmregionsaustralia.com.au/what-is-natural-resource-management/