The core idea
A population estimate combines observations with a model of detection. Survey length, observation conditions and the overlap between observers help explain what a raw count misses.
1. The achievement and its dates
The Wildlife Institute of India records the assessment as Population Status of River Dolphins in India, 2021–23, with a 2024 report citation. Its public release occurred on 3 March 2025. The official release describes surveys across 28 rivers and eight states and an estimate of 6,327 river dolphins. Collection years, report year and release date refer to different stages of the work.
The WII brochure separates Ganges River Dolphins and the small Indus River Dolphin population in India. It describes independent double-observer surveys, underwater acoustic instruments and statistical abundance estimation. These are government research-report sources, not a peer-reviewed journal article. The important achievement is a documented, coordinated evidence base; the headline number should remain attached to its survey period and method.
Sources: WII: Population Status of River Dolphins in India, 2021–23 ↗ · Prime Minister's Office: report release, 3 March 2025 ↗ · WII: river dolphin population assessment brochure and methods ↗
2. Present does not always mean observed
A dolphin can be within the surveyed stretch but underwater while an observer looks. This is an availability problem. It may surface and still be missed because of glare, distance or divided attention, a perception problem. Counting only what is seen can underestimate abundance. Conversely, repeatedly recording the same animal as several individuals can overestimate it. Good protocols address both missed and duplicate observations.
A hydrophone is an underwater microphone. Acoustic records can complement visual observations, but a sound record is not automatically one new animal. The same individual can produce many sounds, while noise can interfere with detection. Combining methods therefore requires a stated matching and analysis procedure. More sensors improve evidence only when their signals are interpreted according to what they actually measure.
Sources: WII: river dolphin population assessment brochure and methods ↗ · Environment Ministry: second dolphin estimation begins, 17 January 2026 ↗ · USGS: capture-mark-recapture science and imperfect detection ↗
3. Worked example: standardise survey effort
Suppose an illustrative team records 50 animals along 100 kilometres and another records 40 along 50 kilometres. Encounter rates are 50 ÷ 100 = 0.5 and 40 ÷ 50 = 0.8 recorded animals per kilometre. The second team has the lower raw count but higher encounter rate. Comparing counts without distance would reverse the impression of which route produced more observations per unit effort.
Even these rates are not automatically population densities. They can differ because river width, season, visibility, observers or detection probability differ. Dividing by kilometres standardises one part of effort, not every influence. An appropriate field record also describes conditions and the surveyed route. If only easy-to-reach stretches are visited, a precise encounter rate may still be unrepresentative of the river system.
Sources: WII: river dolphin population assessment brochure and methods ↗ · USGS: capture-mark-recapture science and imperfect detection ↗
4. Worked example: two observers and unseen animals
Consider two independent observer lists for a short, closed teaching population. Let n1 be the first list's count, n2 the second and m the number correctly matched in both. A simple Chapman-corrected estimate is N = ((n1 + 1)(n2 + 1)/(m + 1)) − 1. N represents estimated abundance. The overlap carries information: if both lists repeatedly find the same individuals, detection may be relatively high.
For n1 = 9, n2 = 11 and m = 5, N = (10 × 12 ÷ 6) − 1 = 19. The number actually seen by at least one observer is 9 + 11 − 5 = 15. The model therefore estimates four additional unseen individuals. These are illustrative counts. WII's assessment uses richer survey data and also describes Huggins modelling; this small calculation does not reproduce the national estimate.
Seen by either observer is not the whole estimate
| Part of the paper model | Animals |
|---|---|
| Observer 1 only: 9 − 5 | 4 |
| Both observers | 5 |
| Observer 2 only: 11 − 5 | 6 |
| Observed union: 4 + 5 + 6 | 15 |
| Chapman estimate | 19 |
Sources: WII: river dolphin population assessment brochure and methods ↗ · USGS: capture-mark-recapture science and imperfect detection ↗
5. The assumptions are part of the result
Closed means the population does not change through arrival, departure, birth or death during the comparison period. Correct matching means a shared record really is the same animal. Independence means one observer's detection does not simply cause the other's detection. Similar or appropriately modelled detection probabilities also matter. If conspicuous animals appear on both lists while hidden animals are missed by both, a simple estimate can be biased.
Notice the sensitivity to overlap. Keeping counts 9 and 11 but changing overlap to 9 gives N = (10 × 12 ÷ 10) − 1 = 11. One small data field substantially changes the estimate. This is why training, independent recording and careful reconciliation matter. An equation cannot rescue unreliable observations, and a point estimate should be accompanied by uncertainty and an explanation of model fit.
Sources: USGS: capture-mark-recapture science and imperfect detection ↗ · WII: river dolphin population assessment brochure and methods ↗
6. The January 2026 update
On 17 January 2026, the environment ministry reported the start from Bijnor of the second rangewide estimation of riverine and estuarine dolphins. Its announcement describes hydrophone monitoring and additional coverage, including Irrawaddy dolphins in the Sundarbans and Odisha. This is a dated announcement of a survey starting and expanding its scope. It is not a published new national population result.
When future results cover additional species or river stretches, their grand total should not be compared mechanically with the earlier total. A valid trend analysis first identifies a comparable subset and checks methods, season and detection. If a school surveys ten classrooms one year and twenty the next, a larger count of students does not by itself prove that class sizes increased. Survey coverage and biological change must be separated.
Sources: Environment Ministry: second dolphin estimation begins, 17 January 2026 ↗
7. Make the estimate useful for conservation
An abundance estimate becomes more useful when linked with habitat and pressure measurements. Researchers can examine whether low-flow stretches, barriers or intensive human activity coincide with particular patterns. Coincidence alone does not prove cause, but it helps frame targeted investigations. Repeated comparable monitoring can test whether an intervention is followed by sustained improvement while accounting for other changes.
For a learner, the practical lesson is to ask three linked questions: what area was surveyed, how likely was detection, and what comparison supports the conclusion? A careful estimate can guide action without pretending every animal was seen. Wildlife observation must also respect the animals and local field protocols; the following exercise uses paper records and requires no approach to dolphins.
Sources: WII: river dolphin population assessment brochure and methods ↗ · Environment Ministry: second dolphin estimation begins, 17 January 2026 ↗ · USGS: capture-mark-recapture science and imperfect detection ↗
PUT IT INTO PRACTICE
Practice: audit a fictional survey
- Make two lists from paper cards representing 12 fictional animals. Record which cards each observer detects and which occur in both lists.
- Using prepared counts n1 = 7, n2 = 8 and m = 5, calculate the Chapman estimate and the observed union. Keep modelled and observed values separate.
- Compare 18 sightings over 30 km with 24 over 60 km. Calculate encounter rates and name one remaining detection difference.
- Write a short report with survey scope, method, estimate and one assumption. Add a sentence explaining what a larger future survey could change.
Check your understanding
What are the two-list practice answers?
N = (8 × 9 ÷ 6) − 1 = 11. The observed union is 7 + 8 − 5 = 10. The model adds one estimated unseen individual.
Which practice route has the higher encounter rate?
The first: 18/30 = 0.6 per km, compared with 24/60 = 0.4. Its smaller raw count does not imply a lower rate.
Why must observers record independently?
If one simply copies the other's detections, overlap no longer supplies independent information about missed animals.
Does a hydrophone count each sound as a different dolphin?
No. Acoustic detections need interpretation, because one animal may produce many sounds and noise can obscure signals.
What does the January 2026 announcement establish?
It establishes the stated start and expanded scope of the second estimation exercise, not a new completed population total.
Why retain uncertainty when communicating the national estimate?
The result is inferred from incomplete detection and a sampling design. Uncertainty helps readers judge precision and whether later changes are meaningful.