Jennifer Balch Hometown: Tunkhannock, Pennsylvania Advisor: Lisa Curran

The Amazon’s transitional forest extends along the southern edge of the basin, coincident with a rapidly expanding frontier driven by large-scale agribusiness and cattle ranching. With increasing ignition sources and a drier climate, escaped wildfires continually threaten this forest. My dissertation research will examine the effects of recurrent understory fire on transitional forest dynamics and flammability in Mato Grosso, Brazil. In one of the largest, replicated experimental burns in the tropics, I will test the effects of two fire regimes. Three experimental blocks (1500 x 1000 m) of intact transitional forest are established within a large private land holding; each block is adjacent to a cattle pasture or agricultural field to simulate escaped fires moving in from the edge. Each replicated 50-hectare block consists of a control, a once-burned, and a twice-burned plot.

My dissertation research will explore the effects of single and recurrent fire on tree mortality, tree growth, and tree regeneration, and resulting changes in forest structure, microclimate, and fuels. My key hypotheses are:   The main predictors of fire probability and intensity will be the quantity and moisture level of fuels and forest interior temperature and relative humidity.   Fire intensity will explain variation in tree mortality, growth, and regeneration patterns.   Forest flammability will increase after an initial burn due to increases in fuel loads from tree mortality, leaf shedding, and potentially live fuels from invading weeds and due to canopy opening which will increase temperature and decrease humidity in the understory.   Due to the increase in forest flammability after an initial fire, fire intensity will increase with a second burn and augment changes in patterns of tree mortality, growth, and regeneration.   An initial fire will cause changes in community composition, favoring species with traits that yield some fire resistance such as thicker bark, but a second, more intense fire will absolve species-specific differences.   Recurrent fires will contribute substantial amounts of carbon to the atmosphere through initial biomass combustion and delayed tree mortality, relative to the amounts of carbon assimilated post-fire in new growth and regeneration.

Using an innovative approach – measuring pre-burn forest properties, quantifying the burn itself, and monitoring the forest’s response – I will test these hypotheses. Flame height, width, and rate of spread are measured at set points along firelines in order to quantify the fire disturbance. Temperature and relative humidity are measured with psychrometers directly before firelines are set. Maximum temperature of the fire is also measured at 2 cm below ground, at the soil surface, and 1 m above ground with Tempilaq heat-senstive paints. Fuel loads and moisture content are measured pre-fire using Brown’s transect method and remaining fuels measured along the same transects post-fire allow calculation of fuel combustion. Temperature and relative humidity are measured continuously (starting 1 month before the burn) with dataloggers at random points throughout the blocks.

A nested sampling design is used to monitor mortality of tree seedlings (? 1 cm dbh), saplings (1-5 cm dbh), sub-canopy trees (5-10 cm dbh), mid-sized canopy trees (10-20 cm dbh), and large canopy trees (? 20 cm dbh). All trees ? 40 cm dbh are monitored throughout the blocks. Trees are tagged, mapped, and identified to species pre-fire and censused one year after each experimental fire. The rate of growth of a subset of trees ? 10 cm dbh is monitored bi-monthly with dendrometer bands. Seedling density in regeneration plots (1 x 0.5 m) is quantified three times: immediately before fire, after fire, and during the wet season. Changes in canopy openness are quantified with hemispherical photographs and LICOR-2000, and through litter fall captured in litter traps (80 x 60 cm) on the ground.

This transitional forest is predicted to undergo continued and extensive land cover change. Fire will be an essential tool in the conversion of forests to other land uses and consequently ignition sources will increase at forest edges. As little is known about how this transitional forest will respond to increasing fire frequency and given the expectation that it will be sensitive to change as a boundary ecotone, it is imperative to understand how recurrent fires will affect forest dynamics and future flammability.