Thesis (M.A.)--Boston University / The primary objective of this research is to test and validate a geometric-optical bidirectional reflectance canopy model developed by Li and Strahler, with respect to actual forest canopy reflectance measurments. This model treats forest canopies as scenes of discrete, three dimensional objects that are illuminated and viewed from different positions in the hemisphere. The shapes of the objects, their count densities and patterns of placement are the driving variables, and they condition the mixture of sunlit and shaded objects and background that is observed from a particular viewing direction, given a direction of illumination. This mixture, in turn, controls the brightness apparent to an observer or a radiometric instrument. The Advanced Solid-State Array Spectroradiometer (ASAS) is chosen to be the sensor having the ability of collecting measurements at various look angles and its imaged reflectance was used to validate the model. The modelled BRF's were compared to actual ASAS measured BRF's in sites with different canopy structures and densities. The comparision revealed execellent match between the modelled and measured reflectance, and great ability of the model in predicting the shape and magnitude of the BRDF, in almost all the sites investigated. It is concluded that the geometric optics approach provided a good way to model the bidirectional reflectance distribution function of natural vegetation canopies, that captures the most important features exhibited by bidirectional measurements of such canopies. Further modifications have been suggested that will improve the predicted BRF's, and yield better results. [TRUNCATED]

Quacking aspen (Populus tremuloides Michx.) is an iconic species in western United States that offers multiple ecosystem services, including carbon sequestration. A shift in forest cover towards coniferous species due to natural succession, land management practices, or climate change may modify soil organic carbon (SOC) dynamics and CO2 emissions. The objectives of this study were to: (i) assess the effects of overstory composition on SOC storage and stability across the aspen-conifer ecotone, (ii) use Fourier transform infrared spectroscopy attenuated total reflectance (FTIR-ATR) to assess whether SOC storage is associated with preferential adsorption of certain organic molecules to the mineral surfaces, and (iii) develop models using near-infrared reflectance spectroscopy (NIRS) to predict aspen- and conifer-derived SOC concentration. Mineral soils (0 – 15 cm) were sampled in pure and mixed aspen and conifer stands in Utah and subjected to physical fractionation to characterize SOC stability (i.e., SOC protected against microbial decomposition), long term laboratory incubations (i.e., SOC decomposability), and hot water extractions (i.e., SOC solubility). Vegetation cover had no effect on SOC storage (47.0 ± 16.5 Mg C ha−1), SOC decomposability (cumulative released CO2-C of 93.2 ± 65.4 g C g−1 C), SOC solubility (9.8 ± 7.2 mg C g−1 C). Mineral-associated SOC (MoM) content was higher under aspen (31.2 ± 15.1 Mg C ha-1) than under mixed (25.7 ± 8.8 Mg C ha−1) and conifer cover (22.8 ± 9.0 Mg C ha−1), indicating that aspen favors long-term SOC storage. FTIR-ATR spectral analysis indicated that higher MoM content under aspen is not due to higher concentration of recalcitrant compounds (e.g., aliphatic and aromatic C), but rather to stabilization of simple molecules (e.g., polysaccharides) of plant or microbial origin. NIRS models performed well during calibration-validation stage (ratio of standard deviation of reference values to standard error of prediction (RPD) ≥ 2). However, model performance decreased during independent validation (RPD = 1.2 – 1.6), probably due to the influence of soil texture, mineralogy, understory vegetation, and land history on SOC spectra. Further improvement of NIRS models could provide insight on SOC dynamics under potential conifer encroachment in semiarid montane forests.

Cette étude propose une analyse intégrée de biomarqueurs moléculaires de sols, depuis leur site de production jusqu'à leur archivage sédimentaire lacustre. Un inventaire des lipides neutres est d'abord réalisé sur des sols d'usages contrastés : prairies/pâtures et forêts. Globalement, cette approche permet de distinguer (1) des composés linéaires ubiquistes, et (2) des composés (poly-)cycliques, généralement spécifiques. Parmi ces derniers figurent les acétates de triterpényle et les méthoxyserratènes, respectivement biomarqueurs d'Astéracées et de Pinacées. La persistance de l'empreinte moléculaire d'un ancien usage des sols est aussi démontrée. Ensuite, l'analyse multi-proxy d'une carotte sédimentaire couvrant les 6700 dernières années révèle l'impact prépondérant des activités humaines sur le fonctionnement hydrologique du lac, depuis l'époque gallo-romaine. Les assemblages moléculaires précédemment définis dans les sols sont globalement retrouvés, associés à un nouveau biomarqueur spécifique du chanvre, d'intérêt paléoenvironnemental. Une tentative de reconstitution des anciennes occupations des sols apparaît conforme aux données historiques et paléoenvironnementales antérieures. / This study proposes an integrated analysis of molecular biomarkers of landuse, from their genesis in soils to their archiving in the lacustrine sediment. An inventory of neutral lipids is realized on soils of two contrasted landuses: grassland/pasture and forest. This approach globally allows to distinguish (1) ubiquist linear compounds and (2) (poly-)cyclic compounds, generally specific. Among these ones, triterpenyl acetates and methoxyserratenes are detected and are respectively biomarkers of Asteraceae and Pinaceae. The persistence of a molecular imprint of an ancient landuse is also demonstrated. Then, the multi-proxy analysis of a sedimentary core, covering the last 6700 years, shows the prominent impact of human activities on the hydrologic functioning of the lake since the gallo-roman period. Molecular assemblages previously defined in soils are globally detected in sediments, associated with a new specific biomarker of hemp, of palaeoenvironmental interest. Hypotheses for the reconstitution of past landuses appear to be consistent with previous historical and palaeoenvironmental data.

Management of natural landscapes requires knowledge of key disturbance processes and their effects. Fire and forest histories provide valuable insight into how fire and vegetation varied and interacted in the past. I constructed multi-century fire chronologies for 10 sites on six mountain ranges representative of the eastern Great Basin (USA), a region in which historic fire information was lacking. I also constructed tree recruitment chronologies for two sites. I use these chronologies to address three research foci. First, using fire-scar data from four heterogeneous sites, I assert that mean fire interval (MFI) values calculated from composite chronologies provide suitable estimates of point MFI (PMFI) when sample area size is ≈&frac; ha. I also suggest that MFI values for single trees can be used to estimate PMFI after applying a correction factor. Next, I infer climate effects on regional fire patterns using 10 site chronologies and tree-ring-based indices of drought and of El Niño Southern Oscillation (ENSO) and Pacific Decadal Oscillation ([PDO), Pacific Ocean surface temperature variability known to affect North American climate. Regional fire years (≥33% of recording sites) were synchronized by wet-dry cycles where the probability of occurrence was highest in the first year of drought following a wet phase and lowest when climate conditions transitioned from dry to wet. Regional fire probability was highest when ENSO and PDO were negative (Southwest pattern). Local fire years occurred under a broad range of conditions. Fire seasonality was bimodal with early and late-season fires dominant. I imply that Native American burning practices were responsible for differences in historic and modern fire seasonality. Lastly, I assess fire regime and tree recruitment variability within two fire-sheds. PMFI varied more than 10-fold within each site. A mixed-severity regime was dominant. A majority (>60%) of fires were small (<10 ha) but together accounted for a minor proportion of area burned. Recruitment pulses varied spatially from stand to landscape-scales and were often synchronous with multi-decade, fire-quiescent periods. I recommend that management strategies employ fire and fire-surrogate treatments to restore disturbance processes to these and similar landscapes at spatial and temporal scales consistent with the historic record.