# Monitoring Amur tiger populations: characteristics of track surveys in snow

> How do you count some of the rarest big cats on Earth across a vast snowy forest? Using five winters of foot surveys in Russia’s Sikhote-Alin reserve, researchers found tiger tracks stayed readable for about a week in mid-winter but only two days in March. They worked out how long survey routes should be, how many are needed, and why counting tracks beats simply noting whether tigers were there.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2002-hayward-miquelle-monitoring-amur-tiger-populations/

## Citation and access

- **Authors:** Gregory D. Hayward, Dale G. Miquelle, Evgeny N. Smirnov and Chris Nations
- **Year published:** 2002
- **Journal or source:** Wildlife Society Bulletin
- **Full citation:** Hayward, G. D., Miquelle, D. G., Smirnov, E. N. and Nations, C. 2002. Monitoring Amur tiger populations: characteristics of track surveys in snow. Wildlife Society Bulletin 30(4): 1150-1159.
- **Access status:** Publisher link only
- **Where to find it:** https://www.jstor.org/action/doBasicSearch?Query=%22Monitoring+Amur+tiger+populations%3A+characteristics+of+track+surveys+in+snow%22

## Tags

- **Species:** Mammals > Tiger
- **Region:** Asia > Russia
- **Tracking:** Snow tracks; Spoor age estimation; Track counts; Track measurements; Tracking; Tracks; Tracks and signs identification; Trailing
- **Methods and technology:** Analysis of variance (ANOVA); Linear regression; Population trends; Power analysis; Simulation
- **Data type:** Modelling or theoretical; New field data
- **Observer qualification:** Not reported
- **Discipline:** Population trends; Wildlife tracking
- **Year published:** 2002

## Detailed keywords

- **Species:** Studied: Amur tiger (Panthera tigris altaica). Tree species named in the study area description: Mongolian oak (Quercus mongolica), Korean pine (Pinus koraiensis), larch (Larix komarovii), birches (Betula costata, B. lanata and others), basswood (Tilia amurensis), Khingan fir (Abies nephrolepis) and Jeddo spruce (Picea ajanensis).
- **Location and habitat:** In and near Sikhote-Alin State Biosphere Zapovednik (a strictly protected reserve), Primorski Krai, Russian Far East, Russia; 13 river basins on the coastal side of the Sikhote-Alin Divide, facing the Sea of Japan. Mostly deciduous forest dominated by Mongolian oak after past fires, with mixed deciduous and conifer forest inland and higher up. Tiger range in Russia covers nearly 200,000 km², at an estimated 0.1 to 1.0 tigers per 100 km².
- **Topics:** winter track counts in snow; track index (tiger crossings divided by route length times days since snow); track longevity and causes of track degradation (fresh snow, wind, ice overflow, melt-out); front pad width; track freshness estimated to the nearest 24 hours; route length; days since last snow; zero counts; repeated crossings by one tiger; statistical power to detect population trends; Type I error (“false alarms”); track counts versus presence-absence surveys; the untested link between a track index and tiger abundance; design of a long-term monitoring program; the traditional Russian tiger census.
- **Statistics or models:** Track index standardised by route length and days since snow; F test comparing track longevity across months; standard deviation and coefficient of variation by route length and time since snow; resampling field routes into 5,000 simulated combined routes per length category; a simulated 60 x 60 km landscape with random tiger trails and survey routes (2,000 iterations for each of 25 route lengths) to model zero counts; linear and negative exponential regression of the proportion of zero counts; Monte Carlo power simulations (10,000 runs, program MONITOR 6.2) using linear regression of log-transformed track indices over 5 years, one-tailed tests, alpha 0.05 to 0.20. Possible inconsistencies: (1) the text says snowfall caused 82% of track degradation in January and February, but Table 1 gives 74 of 96 tracks (77%); (2) the paper reports 538 surveys in the Results, 434 in the abstract and Table 4, and 427 in the variance and power analyses (7 removed for repeated crossings), while the sample sizes in Table 5 add up to 533 by route length and 437 by time since snow, and Table 4’s cells add up to 435; (3) the abstract and discussion give the power result for routes 12 to 15 km long, while the recommended average route length is 10 to 15 km.
- **Data type:** New field data: foot track surveys in winter (January to March), 1995 to 1999, plus repeated checks of individual tiger tracks every second day to measure track longevity. Simulations built on these field data.

## Synopsis

### 1. What they did

Amur tigers in the Russian Far East have been counted for about 50 years with large winter track surveys, where experts interpret the number and spread of tracks to give a full census. Different readings of the same track data have led to fierce debate about how many tigers there really are, and no standard method or measure of error was ever used. The authors wanted a monitoring method that was statistically sound, workable over a vast forested range with very few tigers, and in keeping with the Russian tradition of winter track counts.

Working in and near Sikhote-Alin Zapovednik, a strictly protected reserve, from 1995 to 1999, they:

- Measured how long tracks last. Every second day in January, February and March, they revisited individual tiger tracks and recorded how long the width of the front pad could still be measured, and what ruined the track (fresh snow, wind, ice overflow or melting).

- Walked survey routes. Teams of 5 to 20 observers walked set routes on foot. For each survey they recorded route length, days since the last snow, snow depth, forest type, and the number of tiger trails that crossed the route. For each trail they measured front pad width, estimated freshness to the nearest 24 hours, and noted the slope aspect.

- Built a track index. Only route length and days since snow were clearly related to how many tracks were found, so the index was tiger crossings divided by (route length x days since snow).

- Tested survey designs. Using the field data and computer simulations, they asked how route length affects the variability of the index and the share of routes with no tiger tracks, and how route length, number of routes and the accepted error rate affect the power to detect a population trend over 5 years.

### 2. What they found

The surveys

- They carried out 538 surveys in 13 river basins. Routes averaged 7.4 km and were walked an average of 7.7 days after snow.

- Tiger tracks were found on 42.2% of routes. Of routes with tracks, 51% had only one set and 75% had three or fewer.

How long tracks last

- Tracks could be measured for 8.7 days on average in January and 7.8 days in February, but only 2.1 days in March.

- In March, warm weather melted most tracks. In January and February, fresh snow was the most common cause (82% according to the text). Wind played a minor role, mostly before March, when cold snow blew more easily.

Route length and days since snow

- The index varied less on longer routes and when more days had passed since snow.

- Simulations showed a dramatic drop in variability between the shortest routes (under 3 km) and routes of 3 to 6 km, then smaller and smaller gains. Very long routes bring only marginal improvement for a great deal of effort.

- Longer routes and more days since snow both meant fewer routes with no tiger tracks. Field data showed a straight-line decline, not the curved pattern the simulation predicted.

Power to detect a trend

- A single route had little power to detect any change.

- Ten routes, each surveyed twice a year for 5 years, gave about 89% power to detect a 10% annual decline when accepting a 20% chance of a false alarm. Twenty routes gave about 98%. Going from 3 to 10 routes gave the biggest gain.

- Even 20 routes could not reliably detect a 5% annual change (about 75% power).

- Stricter error rates (below 0.15) gave unacceptably low power for every trend size tested.

### 3. What they concluded

- Track counts could be part of a system to monitor Amur tigers, as long as changes in track counts truly reflect changes in tiger numbers. That critical link has not been tested.

- A system of 10 to 20 routes, 12 to 15 km long, each surveyed twice a year, could give over 80% power to detect a 10% annual decline. They recommend routes averaging 10 to 15 km, because in deep-snow years a field worker probably cannot cover more than 15 km.

- They suggest surveying 6 to 8 days after snow in January and February: few tracks are lost, and variability is lower.

- They recommend no fewer than 10 routes in each count unit. A 10% yearly decline would take an estimated 350 adult tigers down to about 200 in 5 years, so they were uncomfortable designing only to detect larger declines.

- Accepting a 20% false-alarm rate means sometimes concluding tigers are declining when they are not, but stricter rules would delay action on real declines.

- They favour counting tracks over recording only presence or absence, because counts give more power to detect declines, and presence-absence data can still be taken from the same records.

- Monitoring based on this work (and other data) had begun in 16 count units averaging over 1,600 km² across tiger range in Russia.

### 4. How tracking was or was not used

The whole study is built on tracks in snow. Observers identified tiger tracks crossing their routes, counted the crossings, measured the width of the front pad, and estimated the age of each track to the nearest 24 hours. They also followed individual tracks over days to see how long they stayed measurable and what destroyed them.

Pad width was measured to try to tell individual tigers apart, but the authors could not be certain which tracks belonged to which tiger, so they counted crossings rather than individuals. One tiger crossing a route again and again could inflate the count: 7 surveys were removed because the crossings and the tiger’s snow trail suggested a single animal weaving on and off the path. The authors say future protocols need a standard way to deal with this.

### 5. Tracker or observer certification

Not reported. Surveys used 5 to 20 field observers at a time, and the acknowledgements thank Zapovednik scientists and forest guards who helped collect field data. Nothing is said about their tracking experience or any test of their skill, although the counts, freshness estimates and decisions about repeated crossings all depend on reading tracks correctly.

### 6. Other technologies, skills or innovations

- A simple track index that corrects counts for route length and days since snow.

- Computer simulations of tiger trails and survey routes in a model landscape, and power simulations with the program MONITOR 6.2.

### 7. Implications for conservation (from the paper)

- Track counts can offer statistical rigour while keeping a link to the long-established and politically accepted Russian tiger counts. The same program can also collect information on reproduction, prey, human impact and tiger deaths.

- Alongside monitoring, the relationship between track counts and actual tiger numbers should be tested, for example by running track counts where other methods estimate abundance directly.

- The suggested sample sizes should be seen as minimums, with larger samples in the most important regions for tiger conservation.

- If the statistical assumptions are not met, other analysis methods (permutation or randomisation tests) may be needed.

### 8. Implications for tracking (Original Wisdom's observations)

_These are Original Wisdom's observations, not the authors'._

- The study gives trackers practical numbers for snow: tiger tracks stayed readable for about a week in mid-winter but only about two days in March, when melting took over. Track age matters as much as track presence when designing a count.

- The problem of one animal crossing a route many times shows why trailing and individual identification matter: a count is only as good as the ability to tell how many animals made the tracks. Here, pad width alone was not enough to separate individuals with confidence.

## Cite this summary

Lawrence, K. 2026. Summary of: Monitoring Amur tiger populations: characteristics of track surveys in snow (Hayward and Miquelle et al., 2002). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2002-hayward-miquelle-monitoring-amur-tiger-populations/