Biology · Population Ecology & Life-History Patterns
Dispersal, metapopulations and fragmented habitats
Patch B is empty at one census and occupied again later. The species has returned locally, but that does not tell us how many individuals live across the whole region or whether another patch has lost its population. Spatial ecology keeps the patch identities beside the totals.
Organisms, Environments & Ecological Niches introduces ecological scales, habitat and the distinction between observation and inference. Begin here with the counts and ordinary arithmetic. The mathematics explanations of functions and exponentials are optional preparation for the growth lesson; its multiplier-versus-slope explanation can be read before derivatives. Return to Animal Reproduction, Development & Life Histories and Evolutionary Processes & Adaptation for reproductive outcome stages and the evidence needed to connect demographic differences to selection.
Back to the Biology mapBy the end, you can
Track local and regional population boundaries, distinguish occupancy from abundance, and reason about colonisation and local extinction without assuming guaranteed persistence.
Movement changes one account while conserving another
Dispersal is movement that can connect locations or populations; the movement relevant to a demographic account depends on the species and question. Routine movement within one defined population is not necessarily a change of population membership. When individuals move between local patches, a local arrival and departure can occur without crossing the larger regional boundary.
In the sixty-member example, all three patches are inside one fixed region. A begins with 10, B with 20 and C with 30. Individuals moving A→B, B→C and C→A are lost from one local count and gained by another. The complete regional abundance stays 60 because there are no births, deaths or external boundary crossings.
A local increase therefore need not mean regional reproduction or regional population growth. The conserved identities make that explicit: each starting member occurs once at the ending census. If movement beyond the regional boundary, death or recruitment were added, those would need separate entries in the regional balance.
Within-region movement redistributes sixty identities
Read the diagram in words
Before movement, A/B/C counts are 10/20/30. Three move A→B, five B→C and two C→A. The final counts are 9/18/33, still total 60. All original individual identities occur exactly once at both censuses.
| Local patch | Living members in this patch | Conserved individual identities |
|---|---|---|
| Patch A | 10 | A1, A2, A3, A4, A5, A6, A7, A8, A9, A10 |
| Patch B | 20 | B1, B2, B3, B4, B5, B6, B7, B8, B9, B10, B11, B12, B13, B14, B15, B16, B17, B18, B19, B20 |
| Patch C | 30 | C1, C2, C3, C4, C5, C6, C7, C8, C9, C10, C11, C12, C13, C14, C15, C16, C17, C18, C19, C20, C21, C22, C23, C24, C25, C26, C27, C28, C29, C30 |
All living members inside the region: 60
| Local patch | Living members in this patch | Conserved individual identities |
|---|---|---|
| Patch A | 9 | A4, A5, A6, A7, A8, A9, A10, C1, C2 |
| Patch B | 18 | B6, B7, B8, B9, B10, B11, B12, B13, B14, B15, B16, B17, B18, B19, B20, A1, A2, A3 |
| Patch C | 33 | C3, C4, C5, C6, C7, C8, C9, C10, C11, C12, C13, C14, C15, C16, C17, C18, C19, C20, C21, C22, C23, C24, C25, C26, C27, C28, C29, C30, B1, B2, B3, B4, B5 |
All living members inside the region: 60
| Assigned transfer inside the regional boundary | Living members in this patch | Conserved individual identities |
|---|---|---|
| Patch A → Patch B | 3 | A1, A2, A3 |
| Patch B → Patch C | 5 | B1, B2, B3, B4, B5 |
| Patch C → Patch A | 2 | C1, C2 |
Regional boundary containing all three patches
A local arrival can be a regional redistribution
Original simultaneous within-region transfers of distinct starting individuals. No births, deaths, repeated transfers or crossing of the regional boundary; every identity remains counted once. These are assigned movements, not movement protocols or empirical routes.
Count occupied patches without guessing their populations
Patch occupancy records whether a patch contains members of the species. An occupied patch could have few or many individuals. The number of occupied patches is therefore different from regional abundance. The occupancy model supplies complete presence states, so an empty state is known absence; in a field record a nondetection alone might not establish absence.
Local extinction means that a local population disappears from a patch. Colonisation establishes occupancy in a previously empty patch; recolonisation restores it after earlier local extinction. The physical patch can remain available even when its population disappears. These labels describe changes relative to the chosen local boundary.
Our four fixed patches have occupied identities A/B, then A/C, then B/C, then none. The first three occupancy totals are all 2, while identities change. B undergoes local extinction and later recolonisation. The final zero means regional absence at that census because every patch in this stipulated complete region is empty.
Local turnover can occur before regional disappearance
Read the diagram in words
At censuses zero through three the occupied patch counts are 2, 2, 2, 0. The identities are A/B, A/C, B/C and then none. B disappears locally and is recolonised; A later disappears. Both B and C lose occupancy before the final census.
| Patch A | (0, 0) |
|---|---|
| Patch B | (2, 0) |
| Patch C | (0, 2) |
| Patch D | (2, 2) |
| Assigned census position | Patch A | Patch B | Patch C | Patch D | Number of occupied patches in the region |
|---|---|---|---|---|---|
| 0 | Occupied at this census | Occupied at this census | Unoccupied at this census | Unoccupied at this census | 2 |
| 1 | Occupied at this census | Unoccupied at this census | Occupied at this census | Unoccupied at this census | 2 |
| 2 | Unoccupied at this census | Occupied at this census | Occupied at this census | Unoccupied at this census | 2 |
| 3 | Unoccupied at this census | Unoccupied at this census | Unoccupied at this census | Unoccupied at this census | 0 |
| Assigned census position | Persistent patch identity | Local turnover can occur before regional disappearance |
|---|---|---|
| 1 | Patch B → Patch B | Local occupancy becomes absent |
| 1 | Patch A → Patch C | Arrival establishes occupancy in a previously empty patch |
| 2 | Patch A → Patch A | Local occupancy becomes absent |
| 2 | Patch C → Patch B | Occupancy returns after an earlier local extinction |
| 3 | Patch B → Patch B | Local occupancy becomes absent |
| 3 | Patch C → Patch C | Local occupancy becomes absent |
No patch remains occupied at the final census
Turnover alone does not guarantee regional persistence
Original complete patch-occupancy assignments inside one region; all patches remain physically available. Presence means at least one individual, with no local abundance count or detection uncertainty supplied. No incoming colonisation from outside after regional disappearance, persistence probability, empirical dispersal route or guaranteed rescue.
Persistence is a condition to investigate
A metapopulation is a spatial population system in which local populations are connected by dispersal and local extinction and colonisation can matter to regional dynamics. Not every patchy distribution necessarily has frequent local turnover; the useful model depends on the timescale and connections being considered.
Colonisation can allow regional presence to outlast the disappearance of an individual local population. Incoming individuals can sometimes reduce a local population’s risk of extinction, a possibility commonly described as a rescue effect. Such mechanisms require arrivals, suitable conditions and demographic consequences; a line between patches does not guarantee that they occur.
Habitat fragmentation can change the amount, arrangement and connections of suitable habitat. Patch area, isolation, surrounding conditions and the species’ movement abilities can affect different parts of the spatial account. Equal distances on a map are not necessarily equal opportunities for dispersal. The same landscape can be connected for one organism and difficult to cross for another.
The supplied record contains colonisation yet ends in regional absence. It is a finite construction, not a probability forecast. There is no external source population in the final state, and no rule supplying later recolonisation. Demonstrating long-term persistence would require a complete dynamic model or evidence beyond the fact that some earlier turnover occurred.
Worked example
Track the individuals, then track the occupied sites
For within-region movement, A ends at 10 − 3 + 2 = 9, B at 20 − 5 + 3 = 18 and C at 30 − 2 + 5 = 33. Their sum remains 60. The five B individuals moving to C are distinct from the three A individuals arriving in B; the original identity lists prevent accidental double transfer.
In the separate occupancy record, B changes occupied→empty→occupied→empty. A changes occupied→occupied→empty→empty. C changes empty→occupied→occupied→empty. D stays empty throughout. These are presence assignments, not local counts 9, 18, 33 from the other example.
Two occupied patches at successive censuses can conceal local losses and gains. Zero occupied patches at the final census is a different regional outcome. Neither model supplies a colonisation probability, a species-independent corridor rule or a guarantee of recovery.
Check your understanding
Three individuals leave patch A for patch B, both inside the same complete regional boundary. Does regional abundance fall by three?
Show reasoning
No. A loses three and B gains three, so the movements cancel in the regional count. Regional change would require births, deaths or crossing of the regional boundary.
The occupancy total remains two while occupied identities change from A/B to A/C. Has the spatial state remained unchanged?
Show reasoning
No. B has become empty and C occupied. The total hides turnover of local populations. It also supplies no individual abundance within the occupied patches.
Does recolonisation of B at an earlier census guarantee regional persistence in the supplied four-patch record?
Show reasoning
No. B and C both become empty before the final census, leaving no occupied patch. An earlier return demonstrates that event only; later persistence depends on further conditions not guaranteed by the record.
What to carry forward
Keep local identities, regional boundaries and the counted variable visible when interpreting dispersal and spatial persistence.
References
Resources supporting this lesson’s explanations and examples.
- Metapopulation dynamics (Nature 396:41–49, 1998; doi: 10.1038/23876)Ilkka HanskiResearch abstract
- Biology 2e (2018; chapter 45, sections 45.1–45.4)Mary Ann Clark, Matthew Douglas and Jung ChoiTextbook chapters
- Turnover Rates in Insular Biogeography: Effect of Immigration on Extinction (Ecology 58(2):445–449, 1977; doi: 10.2307/1935620)James H. Brown and Astrid Kodric-BrownResearch abstract