Medium-resolution Phenology and Forest Productivity

Appalachian Laboratory researchers collecting tree coresIn recent decades, a trend toward a longer growing season (the number of days between spring onset and autumn offset of greenness) has been observed in some regions of the world.  This phenological change is a response to human-induced changes in both climate and atmospheric chemistry. In turn, changes in growing season length cause impacts to water and carbon cycling in temperate forests.  In fact, observations suggest that a longer growing season has led to increased plant growth in the northern hemisphere, which could create a negative feedback to increasing atmospheric carbon dioxide and climate change.  Therefore, improving our understanding of the relationship between phenological change and forest productivity is important for understanding forest trees and their relationship with global climate.


Over the past 30 years, deciduous forests in the eastern U.S. have been exhibiting contrasting trends in the onset of spring greenness.  Mid-Atlantic forests have demonstrated advancing trends in spring phenological timing while southern Appalachian forests have seen delayed onset of spring greenness. Within each of these regions there is considerable variability in growing season length and forest productivity; from warm wet sites to cool dry sites and everything in between, each forest stand is different. This contrast, both between and within regions, provides a unique opportunity to investigate the importance of phenology and site characteristics in determining the response of tree productivity to a change in growing season length.  Using a novel combination of field and remote-sensing observations, the project synergistically uses phenology observations at both coarse- and medium-resolution, LiDAR observations of canopy complexity, and tree ring records of carbon and nitrogen isotopes to understand the influence of growing season length on tree productivity, measured as tree BAI (annual basal-area increment). 

Preliminary results from Prince William Forest
Preliminary medium-resolution average phenology, LiDAR vegetation structure, and tree-ring data from four sites in Prince William National Forest (one of our northern sites) show that sites with a longer growing season (C & D) exhibited larger BAI over the past 30 years than those with a shorter growing season (A & B before 1990). Further, sites with a longer growing season display a trend toward increased BAI through time, potentially representing a response to a warmer, longer growing season. However, we also saw that sites with a shorter growing season experienced increasingly negative d13C, representing successively more water limitation.

Study Regions

The project focuses on a comparative analysis between regions exhibiting advancing and delayed spring onset of greenness, represented by National Parks and other public forests in the mid-Atlantic and southern Appalachian regions.  The capital region national parks, including Prince William Forest Park, Harpers Ferry National Park, and Catoctin Mountain National Park, encompass all the major phenological gradients in the region.  In the southern Appalachian highlands, the primary park is Great Smoky Mountain National Park.  The native land cover of the study region is primarily deciduous forest, roughly 80% of which fall into two forest categories:  Oak-Hickory and Maple-beech-birch.  Several dendrochronologically important species are found extensively across both regions and are our target species for tree coring:  white oak (Quercus alba), red oak (Quercus rubra), and tulip poplar (Liriodendron tulipifera).


We will be using a novel combination of field and remote-sensing observations to understand the influence of growing season length on tree productivity, measured as tree BAI.  Because remote sensing data are widely considered to provide a consistent source for phenological measurements across space and time, one of our primary tasks will be to assemble coarse- and medium-resolution phenology observations for each region.  The methods utilize multi-temporal observations of canopy greenness at two spatial scales (8km AVHRR GIMMs NDVI [coarse] and 30m Landsat observations of vegetation from spectral mixture analysis [medium]) to model greenness phenology.  

Because canopy complexity has been hypothesized to influence phenological timing in the spring and autumn, LiDAR is being used to quantify canopy height and complexity for areas within each study region.  In our study regions, understory vegetation with differing phenology, including evergreen species and invasive species, are common.  Quantifying canopy height and complexity will help determine how understory vegetation might influence the phenology signal observed by remote sensing platforms.

To further understand the forest dynamics related to changing phenology, we are also making multi-temporal measurements of leaf nitrogen concentration, tree ring width, d13C, and d15N at sites spanning phenology gradients.  It has been observed that trees that are not resource limited are typically more productive and allocate relatively more carbon to wood (i.e. they produce larger rings) while water-limited trees produce thinner rings. Tree-ring δ13C can be used to reconstruct intra- and inter-annual variations in plant moisture status, while the nitrogen isotopic composition of tree cores can indicate nitrogen availability.

Following field collection, each ring is sliced into earlywood and latewood sections.  The latewood portions, which reflect summer growth, are analyzed for carbon concentration and δ13C and nitrogen concentration and δ15N, respectively, using a Carlo Erba NC2500 elemental analyzer interfaced with a Thermo Delta V+ isotope ratio mass spectrometer (IRMS) at the Central Appalachians Stable Isotope Facility (CASIF).

An example tree core
Example of a tree core from which earlywood and latewood are sampled for isotope analysis.


At a subset of sites with high accessibility that exhibit both low and high greendown, we are also conducting repeat measurements reflectance along with foliar nitrogen, δ15N, and δ13C at bi-weekly intervals to help determine the extent to which leaf structure and chemistry regulate greendown.

Project PI's:  Andrew Elmore, David Nelson

This website was prepared by the University of Maryland Center for Environmental Science under award (#NNX12AK17G) from the National Aeronautics and Space Administration. The statements, findings, conclusions and recommendations are those of the author's) and do not necessarily reflect the views of National Aeronautics and Space Administration.