QUANTIFYING FIRE REGIMES BASED ON TREE-RING METHODS

A. Fire-scar Methods

1. Recognizing fire scars

Scar initiation: heating of the cambium temperature to lethal temperatures

consumption of the bark by fire (sometimes)

subsequent sloughing off of the bark after the cambium dies

scar is usually on leeward side of tree

(The scar is usually on the uphill side because fire tends to spread uphill-- tree slows the spread of the fire so that the longer residence time of the leeward flame vortex heats the cambium more on the leeward side)

Susceptibility to fire scarring varies with:

bark thickness (i.e. probable variation by tree age/size and species)

stem diameter (small diameters may not slow the fire down enough to allow enough heating for scar formation; also seedlings may not survive)

previous scarring (once scarred, then subsequent low intensity fires are more likely to be recorded due to exposure of sapwood, perhaps with flammable resin)–“fire-scar susceptible trees” or “recorder trees”.

topographic position, wind, and surrounding fuel conditions

Not all fires are recorded by fire scars!

Recognizing fire scars – triangular shaped, at base of tree, possibly charcoal on twigs, usually on multiple trees, usually on uphill side

Other types of scars –

Lightning– long narrow strip, sometimes spiraling

Humans– blaze marks; bark stripping (oval to rectangle, tool marks)

Bears– peeled bark, claw marks; usually not synchronous on multiple trees

Porcupines, squirrels– not triangular

Voles– chew at the base only

Deer, elk– antler rubbing at about 1 m above the ground

Beetle “patch scars” – usually not triangular

Beetle strip attacks– beetle holes in bark, beetle galleries, blue stain fungi, retained bark on scar face, lack of char, rectangular rather than triangular

Basal scars from fungi (e.g. Armillaria)-- scar associated with root decay

Windthrow scrape scars– gouged or damaged sapwood, broken branches

Sunscald– long narrow scars, usually not to the base, on edge of sudden gap

Scarlets– possibly due to blowing ice or sand

Always make detailed description of scar position (height, azimuth), shape, etc. to rule out non-fire causes (e.g. protected from blowing ice).

Some potential problems–

The fire scar may heal completely (i.e. some or all interior scars)

First scar may not be due to fire, but the surface will be charred by subsequent scars

2. Dating fire scars

Obtain a master tree-ring chronology for identification of marker years on the fire-scarred cross-sections.

Segregate samples by quality, and start with the best preserved samples with distinct rings and distinct fire-scar tips.

Work with samples from very close neighbors that are likely to share common fire dates.

Identify fire scars from a break or gap along the ring boundary, presence of charcoal on this break, and/or overlapping curvilinear growth over the break.

Tentatively date scars into categories of different confidence levels (e.g., “certain”, + 1 year, + 5 years, etc.)

Revise tentative dates by identifying marker years on the interior side of the tree (i.e. rings towards the bark are likely to be missing). When necessary, measure from the scar towards the interior of the tree and quantitatively cross date (don’t include sections that are too close to a preceding scar).

Note: do not arbitrarily “adjust” dates to synchronize fire dates (e.g. Arno and Sneck 1977); however, in making cross-dating decisions it is appropriate to use subjective judgment in deciding on the correct date.

3. Supplemental evidence of fire dates

a) Postfire cohort ages

This suffers from all the usual problems associated with accurate germination dates and also the problems of:

i) Unknown time lags between fires and tree establishment

ii) The problem of identifying the oldest tree in the cohort.

In fourteen post-fire Ponderosa pine/Douglas fir stands with 100% age data (c. 800 ages) we found:

the largest tree was the oldest (at coring height) in 5 stands: the largest tree was within 5 years of the oldest in 1 stand; in 6 stands the largest tree was 6-15 years younger than the oldest; in 2 stands it was more than 15 years younger than the oldest tree.

Clearly, multiple samples of the largest trees must be taken to estimate stand age.

To estimate the age of the oldest tree in postfire lodgepole stands, sample > 10 of the largest trees in young (c. 100 yrs old) stands and > 15 in old (> 250 yrs) stands.

In combination with precise fire-scar dates, estimates of post-fire cohort ages can be used to map past fire extent.

b) Post-fire releases– highly variable lag times (e.g. up to 12 years in the montane zone in Boulder County).

4. Research design as determined by objectives, forest conditions and fire behavior

Most frequent objectives:

a) “Before and after” comparisons of fire frequency in a fixed area

b) Spatial comparisons of fire frequency for different vegetation types or different land use histories

c) Relationship of fire (frequency, extent) to climatic variation (interannual versus multi-decadal variation)

Forest conditions and fire behavior (surface vs. stand-replacing) constrain options. A continuum from:

a) savanna-like, sparse woodlands of exclusively surface fires (suitable for the “fire interval” approach) to

b) dense flammable forests in which 100% of the trees within the fire perimeter burn in each fire event (completely stand-replacing fires)

How true is either extreme?

c) mixed severity fire regime (also “variable and mixed” or “moderate”) is poorly defined but implies fire events in which there is both a high severity and low severity component (extent of each is not defined)

Focus on the fire interval approach in Ponderosa pine forests (later consider the postfire cohort/stand-origin method).

5. Analysis of fire scar dates

Fire Frequency = number of fires per unit of time in a designated area (the area could be a point, in the case of a single tree)

Fire interval = number of years between successive fire events in a designated area

Composite Fire Interval = interval between successive fire years in a designated area; derived from listing all the fire dates and intervals for an area that might be a small search area or a combination of search areas. CFI is highly dependent on the size of the area (larger areas generally yield more fire years and therefore lower fire intervals).

Master Fire Chronology (or Chart) lists each tree each fire year, beginning of the tree-ring series (i.e. pith date vs. innermost ring) and end of the tree-ring series (i.e. date of outermost year).

Fire rotation = the time required to burn an area equal to the area of the study area. This requires knowledge of fire intervals and the area burned in each fire. Fire rotation is normally not available for ponderosa pine forests because of the lack of data on extent of past fires.

Mean fire interval = MFI (the Weibull Median Probabily Interval which is preferred for skewed distributions, but most fire history studies use MFI).

Two types of MFI: composite MFI vs. point MFI.

Composite MFI– based on fire intervals occurring within a designated area. Usually computed from only scar-to-scar (SS) intervals, but has also sometimes includes the intervals from scar-to-present (SP) or from tree origin date to first scar (OS).

Advantages of composite MFI–

1) increases the chances of dating fires not recorded by all trees (Tucson school);

2) it increases the number of fire years to allow analysis of temporal trends.

Disadvantages of composite MFI–

1) It is highly dependent on size of the search area;

2) It does not discriminate between small versus large fires or even between overlapping and non-overlapping fires (but this can be partially resolved by identifying years of high % fire scars or by mapping individual trees).

3) When fire years are infrequent, MFI changes greatly when only slight changes are made in the definitions of periods such as Native American, Euro-American Settlement, and Fire Suppression Periods.

4) Composite MFI is also dependent on the number of trees sampled (the number of detected fires rises steeply until a sample size of c. 10 trees has been used, but this is only a rule of thumb).

Point MFI– This is the mean individual-tree MFI which is computed by determining 1) the mean interval between scars on each tree, and then 2) then averaging those means (i.e. it is the mean of all the means from the individual trees). Only multiple-scarred trees enter into the computations.

Advantage of point MFI: it is not dependent on the size of the sample area.

Disadvantage of point MFI: on cat faces, some scar dates may be lost from erosion, weathering, overburning, decay, etc.

Use of either composite or point MFIs in statistical tests is problematic due to:

i) skewed distributions of fire intervals (there is a lower limit of 1 year but no upper limit to fire intervals)

ii) fire is spatially autocorrelated

B. Post-fire cohort (i.e. Stand-origin) method

1. Procedure:

Heinselman (1975) produced a "stand‑origin" map for Boundary Waters Canoe Area in Minnesota based on cohort ages of postfire stands.

Stand‑origin map = time‑since fire map = time‑since most recent fire map.

Patches (polygons) that potentially represent past fires must be mapped from air photos, field checked for homogeneity within each patch (i.e. are there enclosures of older/younger patches?), and dated for fire evidence.

Field samples usually consist of a sample (15 to 30 minimum) of the ages of the oldest appearing trees; fire scars are sampled to link the approximate age of the cohort to a precise fire‑scar date.

2. Field constraints/problems:

Requires stand‑replacing fires with persistent sharp boundaries between fires (e.g. dense forests, dense shrublands).

The usual problems of determining total tree age (germination date), lag between fire and germination, and of identifying the oldest tree in a postfire cohort.

Precise dating of past fires still requires fire‑scar dates which may not be available in all vegetation types.

In addition, boundaries between different cohorts may not be distinct and usually require intensive field observation and field data collection.

Other disturbances potentially can be confused with fire as the originating agent, especially after > 200 years of stand development.

3. Use of stand-origin maps

a. To interpret origins of existing forest in relation to fire events

b. To reconstruct extent of all previous fires (overburning of evidence requires subjective decisions about past fire boundaries)