Forests ecosystems cover about 30% of the Earth’s land surface, corresponding to an area of roughly 42 million km2 globally. Forests play an important role in the global carbon cycle by exchanging carbon dioxide (CO2) with the atmosphere. Annually, forests act to effectively sequester large amounts of anthropogenically-emitted CO2 from the atmosphere through photosynthetic processes. Through the unparalleled increase of CO2 emissions over the past century and the subsequent climatic inconsistencies due to global climate change, the carbon sink-capacity of the world’s forests remains uncertain. Furthermore, since increasing temperatures have been shown to extend the vegetative growing season in forests, phenological responses to this change are of particular interest. In an effort to effectively assess the future carbon sequestration potential of forests, a better understanding of the climatic controls on phenology, and its influence on carbon processes, is needed. The eddy covariance (EC) technique is a stand-level, in-situ, method used widely to assess the net CO2 exchange across the canopy-atmosphere interface. Together with meteorological data, the sequestration of CO2 and the subsequent ecosystem productivity can be quantified over various time scales (half-hours to decades). This dissertation reports results from field observations of EC measured fluxes that were used to study the climatic impacts on forest phenology and the resulting carbon dynamics in southern Ontario, Canada. The study sites, part of the Turkey Point Observatory, consisted of two similarly-aged, temperate, North American forests growing under similar climatic and edaphic conditions: the 80-year old (in 2019) white pine plantation (coniferous evergreen) and 90+ year-old, naturally-regenerated, white oak (deciduous broadleaf) forest. These forests were studied from 2012 to 2017, using the EC technique, digital phenological cameras, and remote-sensing measurements. At the deciduous broadleaf forest, mid-summer (July and August) meteorological conditions were the key period in determining the annual carbon sink-strength of the site, acting to regulate the interannual variability in carbon uptake. The forest experienced higher net ecosystem productivity (+NEP; carbon sink) when soil temperatures ranged from 15 to 20°C and vapor pressure deficit was 0.7 and 1.2 kPa. From 2012 to 2016, the forest remained a net annual sink, with mean NEP of 206 ± 92 g C m-2 yr-1, similar to that of other North American deciduous forests. Spring and autumn phenological transition dates were calculated for each year (2012 to 2017) from measured EC data and digital camera greenness indices. The timing of spring and autumn transition dates were impacted by seasonal changes in air temperature and other meteorological variables. Contrary to past studies, an earlier growing season start did not equate to increased annual carbon uptake. In autumn, a later end to the deciduous forest growing season negatively impacted the net carbon uptake of the forest, as ecosystem respiration (RE) outweighed the gains of photosynthesis. The digital camera indices failed to capture the peak dates of photosynthesis, but accurately measured the spring and autumn transition dates, which may be useful in future remote sensing applications. A comparison of the two forests from 2012 to 2017 found the coniferous forest to have higher but more variable annual NEP (218 ± 109 g C m-2 yr-1) compared to that of the deciduous broadleaf forest (200 ± 83 g C m-2 yr-1). Similarly, the mean annual evapotranspiration (ET) was higher (442 ± 33 mm yr-1) at the coniferous forest compared to that of the broadleaf forest (388 ± 34 mm yr-1). The greatest difference between years resulted from the response to heat and drought. During drought years, deciduous carbon and water fluxes were less sensitive to changes in temperature or water availability compared to the evergreen forest. Carotenoid sensitive vegetative indices and the red-edge chlorophyll index were shown to effectively capture seasonal changes in photosynthesis phenology within both forests via proximal remote sensing measurements during the 2016 growing season. Satellite vegetative indices were highly correlated to EC photosynthesis, but significant interannual variability resulted from either meteorological inputs or the heterogeneous landscapes of the agriculturally-dominated study area. This dissertation improved our understanding of the dynamics of carbon exchange within the northeastern North American deciduous forest ecozone, through the examination of climatic variability and its impact on carbon and phenology. This dissertation also contributed to efforts being made to better evaluate the impact of species composition on carbon dynamics in geographically similar forests. Moreover, the use of the digital phenological camera observations and remote sensing techniques to complement and better understand the fluxes observed with the EC method was innovative and may help other researchers in future studies. / Dissertation / Doctor of Philosophy (PhD)

Air quality management strategies in the US are developed largely from estimates of emissions, some highly uncertain, rather than actual measurements. Improved knowledge based on measurements of real-world emissions is needed to increase the effectiveness of these strategies. Consequently, the objectives of this research were to (1) quantify relationships among urban emissions sources, land use, and demographics, (2) determine the spatial and temporal variability of emissions, and (3) evaluate the accuracy of official emissions estimates. These objectives guided three field campaigns that employed a unique mobile laboratory equipped to measure pollutant fluxes by eddy covariance. The first campaign, conducted in Norfolk, Virginia, represented the first time fluxes of nitrogen oxides (NO_x) were measured by eddy covariance in an urban environment. Fluxes agreed to within 10% of estimates in the National Emissions Inventory (NEI), but were three times higher than those of an inventory used for air quality modeling and planning. Additionally, measured fluxes were correlated with road density and increased development. The second campaign took place in the Tijuana-San Diego border region. Distinct spatial differences in fluxes of carbon dioxide (CO₂), NO_x, and particles were revealed across four sampling locations with the lowest fluxes occurring in a residential neighborhood and the highest ones at a port of entry characterized by heavy motor vehicle traffic. Additionally, observed emissions of NO_x and carbon monoxide were significantly higher than those in emissions inventories, suggesting the need for further refinement of the inventories. The third campaign focused on emissions at a regional airport in Roanoke, Virginia. NOx and particle number emissions indices (EIs) were calculated for aircraft, in terms of grams of pollutant emitted per kilogram of fuel burned. Observed NO_x EIs were ~20% lower than those in an international databank. NO_x EIs from takeoffs were significantly higher than those from taxiing, but relative differences for particle EIs were mixed. Observed NO_x fluxes at the airport agreed to within 25% of estimates derived from the NEI. The results of this research will provide greater knowledge of urban impacts to air quality and will improve associated management strategies through increased accuracy of official emissions estimates. / Ph. D.

No description available.

Boreal old-growth forests are key determinants in the global carbon cycle. It is unknown how the role of persistent old-growth forests will be in the carbon cycle in the face of predicted climatic changes. Eddy-covariance measurements are commonly used to quantify carbon exchange between ecosystems, such as forests, and the atmosphere. Error due to gap-fill method is of particular interest in these datasets. Here we filled a 15-year eddy covariance dataset from the Northern Old-Growth Boreal Black Spruce (Picea mariana) site located near Thompson, in central Manitoba, Canada using four different gap-fill methods. Our objectives were to determine if choice of gap-fill method affected annual NEP and if these errors compounded to even greater differences over the 15-year study period. Most significant differences in NEP among methods occurred from September to December, but variations during the growing season were responsible for most of the annual differences. / October 2016

<div>Halogen chemistry in the Arctic boundary layer catalytically destroys O3 and impacts the chemical lifetimes of hydrocarbons, the HOx-NOx cycle, and atmospheric mercury. While many advances have been made in the last several decades in understanding the sources, sinks, and recycling pathways of halogens in the Arctic there are still many unknowns. Previous studies have shown that Br2, BrCl and Cl2 are produced photochemically in the Arctic tundra snowpack, but the magnitude of this production is still poorly understood. Additionally, while there have been suggestions that the tundra snowpack should also produce I2, there have been no previous measurements of I2 in the Arctic. The lack of measurements of the halogen production capacity of Arctic snowpacks has left the community to rely on one-dimensional modeling to estimate the impact of snowpack-derived halogen chemistry on the Arctic atmosphere. Because modeling is inherently dependent on understanding recycling mechanisms, mixing processes, and sinks this leaves the effect of halogens on atmospheric chemistry in the Arctic highly uncertain.</div><div><br></div><div>This work describes efforts to address these uncertainties through measurements made during two field campaigns in Utqiaġvik (formerly Barrow), Alaska in January – February 2014, and February – May 2016. The first measurements of I2 in the Arctic, both in the snowpack interstitial air, and in the air above the snowpack demonstrate that iodine chemistry is active in the Arctic atmosphere, and that I2 is produced photochemically in the tundra snowpack. The effects of active iodine chemistry on both O3 and bromine chemistry is examined through zero- and one dimensional modeling. The first speciated measurements of snowpack phase iodine reveal that much like previous reports of iodine enriched aerosols, the Arctic snowpack is highly enriched in iodine. Vertical profiles of I- in the snowpack suggest that there is a consistent, non-radiation dependent source of iodine to the Arctic environment. It seems likely that this source is transport of iodine-enriched aerosols from the mid-latitudes. However, unlike the Antarctic, and previous</div><div>observations in the mid-latitudes, most Arctic snowpack phase iodine is inorganic, which may contradict transport from the mid-latitudes as a source. One-dimensional modeling was also utilized, in conjunction with the first vertical profile measurements of Br2 and Cl2 between 1 and</div><div>7 m above the snowpack surface to examine the community’s understanding of recycling mechanisms, mixing, sources, and sinks of halogens in the Arctic Atmosphere.</div>

The terrestrial biosphere impacts considerably on the global carbon cycle. In particular, ecosystems contribute to set off anthropogenic induced fossil fuel emissions and hence decelerate the rise of the atmospheric CO₂ concentration. However, the future net sink strength of an ecosystem will heavily depend on the response of the individual processes to a changing climate. Understanding the makeup of these processes and their interaction with the environment is, therefore, of major importance to develop long-term climate mitigation strategies. Mathematical models are used to predict the fate of carbon in the soil-plant-atmosphere system under changing environmental conditions. However, the underlying processes giving rise to the net carbon balance of an ecosystem are complex and not entirely understood at the canopy level. Therefore, carbon exchange models are characterised by considerable uncertainty rendering the model-based prediction into the future prone to error. Observations of the carbon exchange at the canopy scale can help learning about the dominant processes and hence contribute to reduce the uncertainty associated with model-based predictions. For this reason, a global network of measurement sites has been established that provides long-term observations of the CO₂ exchange between a canopy and the atmosphere along with micrometeorological conditions. These time series, however, suffer from observation uncertainty that, if not characterised, limits their use in ecosystem studies. The general objective of this work is to develop a modelling methodology that synthesises physical process understanding with the information content in canopy scale data as an attempt to overcome the limitations in both carbon exchange models and observations. Similar hybrid modelling approaches have been successfully applied for signal extraction out of noisy time series in environmental engineering. Here, simple process descriptions are used to identify relationships between the carbon exchange and environmental drivers from noisy data. The functional form of these relationships are not prescribed a priori but rather determined directly from the data, ensuring the model complexity to be commensurate with the observations. Therefore, this data-led analysis results in the identification of the processes dominating carbon exchange at the ecosystem scale as reflected in the data. The description of these processes may then lead to robust carbon exchange models that contribute to a faithful prediction of the ecosystem carbon balance. This work presents a number of studies that make use of the developed data-led modelling approach for the analysis and interpretation of net canopy CO₂ flux observations. Given the limited knowledge about the underlying real system, the evaluation of the derived models with synthetic canopy exchange data is introduced as a standard procedure prior to any real data employment. The derived data-led models prove successful in several different applications. First, the data-based nature of the presented methods makes them particularly useful for replacing missing data in the observed time series. The resulting interpolated CO₂ flux observation series can then be analysed with dynamic modelling techniques, or integrated to coarser temporal resolution series for further use e.g., in model evaluation exercises. However, the noise component in these observations interferes with deterministic flux integration in particular when long time periods are considered. Therefore, a method to characterise the uncertainties in the flux observations that uses a semi-parametric stochastic model is introduced in a second study. As a result, an (uncertain) estimate of the annual net carbon exchange of the observed ecosystem can be inferred directly from a statistically consistent integration of the noisy data. For the forest measurement sites analysed, the relative uncertainty for the annual sum did not exceed 11 percent highlighting the value of the data. Based on the same models, a disaggregation of the net CO₂ flux into carbon assimilation and respiration is presented in a third study that allows for the estimation of annual ecosystem carbon uptake and release. These two components can then be further analysed for their separate response to environmental conditions. Finally, a fourth study demonstrates how the results from data-led analyses can be turned into a simple parametric model that is able to predict the carbon exchange of forest ecosystems. Given the global network of measurements available the derived model can now be tested for generality and transferability to other biomes. In summary, this work particularly highlights the potential of the presented data-led methodologies to identify and describe dominant carbon exchange processes at the canopy level contributing to a better understanding of ecosystem functioning. / Der Kohlenstoffhaushalt der Erde wird maßgeblich von der bewachsenen Landoberfläche beeinflusst. Insbesondere tragen terrestrische Ökosysteme dazu bei, den Anstieg der atmosphärischen Kohlenstoffdioxid- (CO₂-) Konzentration durch anthropogen verursachte Emissionen fossiler Brennstoffe zu verlangsamen. Die Intensität der Netto-CO₂-Aufnahme wird allerdings in einem sich verändernden Klima davon abhängen, wie einzelne Prozesse auf Änderungen der sie beeinflussenden Umweltfaktoren reagieren. Fundierte Kenntnisse dieser Prozesse und das Verständnis ihrer Wechselwirkungen mit der Umwelt sind daher für eine erfolgreiche Klimaschutzpolitik von besonderer Bedeutung. Mit Hilfe von mathematischen Modellen können Vorhersagen über den Verbleib des Kohlenstoffs im System Boden-Pflanze-Atmosphäre unter zukünftigen Umweltbedingungen getroffen werden. Die verantwortlichen Prozesse und ihre Wechselwirkungen mit der Umwelt sind jedoch kompliziert und bis heute auf der Ökosystemskala nicht vollkommen verstanden. Entwickelte Modelle und deren Vorhersagen sind deshalb derzeit mit erheblichen Unsicherheiten behaftet. Messungen von CO₂-Austauschflüssen zwischen einem Ökosystem und der Atmosphäre können dabei helfen, Vorgänge besser verstehen zu lernen und die Unsicherheiten in CO₂-Austausch-Modellen zu reduzieren. Allerdings sind auch diese Beobachtungen, wie alle Umweltmessungen, von Unsicherheiten durchsetzt. Ziel dieser Arbeit ist es Methoden zu entwickeln, die physikalisches Prozessverständnis mit dem dennoch großen Informationsgehalt dieser Daten vorteilhaft zu vereinigen. Dabei soll vereinfachtes Prozessverständnis dazu genutzt werden, Zusammenhänge zwischen dem CO₂-Austausch und den umgebenden Umweltbedingungen aus den Beobachtungen abzuleiten. Das Besondere hierbei ist, dass diese Zusammenhänge direkt aus den Daten geschätzt werden, ohne vorher Annahmen über ihre funktionale Form zu machen. Die Daten als Ausgangspunkt der Modellentwicklung zu wählen gewährleistet, dass die Komplexität der Modelle dem Informationsgehalt der Messungen entspricht. Auf diese Weise lassen sich diejenigen Prozesse identifizieren, welche für den CO₂-Austausch mit der Atmosphäre dominant sind. Die gewonnenen Erkenntnisse können dann in robuste CO₂-Austauschmodelle für Ökosysteme überführt werden und zur Vorhersage von Kohlenstoffbilanzen beitragen. In der vorliegenden Arbeit werden diese entwickelten, datenbasierten Methoden zur Analyse und Interpretation von Netto-CO₂-Flüssen eingesetzt. Die erste Studie führt ein datenbasiertes Modell ein, das unvermeidliche Lücken in Messzeitreihen zuverlässig interpoliert. Dies ermöglicht erweiterte Anwendungen der Daten. In einer nächsten Studie wird ein Verfahren vorgestellt, mit dem die Unsicherheiten in den Beobachtungen charakterisiert werden können. Dies ist nötig, um jährliche Kohlenstoffbilanzen von Ökosystemen unter Berücksichtigung der Messungenauigkeiten direkt aus den Daten herzuleiten. Dabei liegt die Unsicherheit in den betrachteten Waldstandorten bei maximal 11% des Jahreswertes. In einer weiteren Studie werden dieselben Modelle genutzt, um die Netto-CO₂-Flüsse in Einzelkomponenten der CO₂-Assimilation und -Abgabe zu bestimmen. Diese Komponenten sowie die Nettobilanz sind zusammen mit ihren Ungenauigkeiten für Vorhersagen über das Kohlenstoffsenkenpotential eines Ökosystems von besonderer Bedeutung und können Abschätzungen des globalen Kohlenstoffhaushaltes maßgeblich unterstützen. Abschließend zeigt die letzte Studie ein Beispiel für die datenbasierte Entwicklung eines Modells, das die dominanten Prozesse des Kohlenstoffaustausches in Waldökosystemen beschreibt und erfolgreich vorhersagen kann. Dies unterstreicht insbesondere das Potenzial des vorgestellten Modellierungsansatzes, vorherrschende Prozesse zu identifizieren, zu beschreiben und damit zum verbesserten Verständnis des CO₂-Austauschs zwischen Ökosystem und Atmosphäre beizutragen.

The exact quantification of site water budget is a necessary precondition for successful and sustainable management of forests, agriculture and water resources. In this study the water balance was investigated at the spatial scale of canopies and at different temporal scales with focus on the monthly time scale. The estimation of the individual water balance components was primarily based on micrometeorological measurement methods. Evapotranspiration was assessed by the eddy-covariance (EC) method, while sap flow measurements were used to estimate transpiration. Interception was assessed by a combination of canopy drip, stem flow and precipitation (gross rainfall) measurements and soil moisture measurements were used to estimate the soil water storage. The combination of different measurement methods and the derivation of water balance components that are not directly measurable e.g. seepage and soil evaporation is a very complex task due to different scales of measurement, measurement uncertainties and the superposition of these effects. The quantification of uncertainties is a core point of the present study. The uncertainties were quantified for water balance component as well as for meteorological variables (e.g. wind speed, temperature, global radiation, net radiation and precipitation) that served as input data in water balance models. Furthermore, the influences of uncertainties were investigated in relation to numerical water balance simulations. Here, both the effects of uncertainties in input data and in reference data were analysed and evaluated. The study addresses three main topics. The first topic was the providing of reference data of evapotranspiration by EC measurements. Here, the processing of EC raw-data was of main concern with focus on the correction of the spectral attenuation. Four different methods of spectral correction were tested and compared. The estimated correction coefficients were significantly different between all methods. However, the effects were small to absolute values on half-hourly time scale. In contrast to half-hour data sets, the method had significant influence to estimated monthly totals of evapotranspiration. The second main topic dealt with the comparison of water balances between a spruce (Picea abies) and a beech (Fagus sylvatica) site. Both sites are located in the Tharandter Wald (Germany). Abiotic conditions are very similar at both sites. Thus, the comparison of both sites offered the opportunity to reveal differences in the water balance due to different dominant tree species. The aim was to estimate and to compare all individual components of the water balance by a combination of the above mentioned measurement methods. A major challenge was to overcome problems due different scales of measurements. Significant differences of the water balances between both sites occurred under untypical weather conditions. However, under typical condition the sites showed a similar behaviour. Here, the importance of involved uncertainties deserved special attention. Results showed that differences in the water balance between sites were blurred by uncertainties. The third main topic dealt with the effects of uncertainties on simulations of water balances with numerical models. These analyses were based on data of three sites (Spruce, Grass and Agricultural site). A kind of Monte-Carlo-Simulation (uncertainty model) was used to simulate effects of measurement uncertainties. Furthermore, the effects of model complexity and the effect of uncertainties in reference data on the evaluation of simulation results were investigated. Results showed that complex water balance models like BROOK90 have the ability to describe the general behaviour and tendencies of a water balance. However, satisfying quantitative results were only reached under typical weather conditions. Under untypical weather e.g. droughts or extreme precipitation, the results significantly differed from actual (measured) values. In contrast to complex models, it was demonstrated that simple Black Box Models (e.g. HPTFs) are not suited for water balance simulations for the three sites tested here. / Die genaue Quantifizierung des Standortswasserhaushalts ist eine notwendige Voraussetzung für eine erfolgreiche und nachhaltige Bewirtschaftung von Wäldern, Äckern und Wasserressourcen. In dieser Studie wurde auf der Raumskala des Bestandes und auf verschieden Zeitskalen, jedoch vorrangig auf Monatsebene, die Wasserbilanz untersucht. Die Bestimmung der einzelnen Wasserbilanzkomponenten erfolgte hauptsächlich mit mikrometeorologischen Messmethoden. Die Eddy- Kovarianz- Methode (EC- Methode) wurde benutzt zur Messung der Evapotranspiration, während Xylem- Flussmessungen angewendet wurden, um die Transpiration zu bestimmen. Die Interzeption wurde aus Messungen des Bestandesniederschlags, des Stammablaufs und des Freilandniederschlags abgeleitet. Messungen der Bodenfeuchte dienten zur Abschätzung des Bodenwasservorrats. Die Kombination verschiedener Messmethoden und die Ableitung von nicht direkt messbaren Wasserhaushaltkomponenten (z.B. Versickerung und Bodenverdunstung) ist eine äußerst komplexe Aufgabe durch verschiedenen Messskalen, Messfehler und die Überlagerung dieser Effekte. Die Quantifizierung von Unsicherheiten ist ein Kernpunkt in dieser Studie. Dabei werden sowohl Unsicherheiten in Wasserhaushaltskomponenten als auch in meteorologischen Größen, welche als Eingangsdaten in Wasserbilanzmodellen dienen (z.B. Windgeschwindigkeit, Temperatur, Globalstrahlung, Nettostrahlung und Niederschlag) quantifiziert. Weiterführend wird der Einfluss von Unsicherheiten im Zusammenhang mit numerischen Wasserbilanzsimulationen untersucht. Dabei wird sowohl die Wirkung von Unsicherheiten in Eingangsdaten als auch in Referenzdaten analysiert und bewertet. Die Studie beinhaltet drei Hauptthemen. Das erste Thema widmet sich der Bereitstellung von Referenzdaten der Evapotranspiration mittels EC- Messungen. Dabei waren die Aufbereitung von EC- Rohdaten und insbesondere die Dämpfungskorrektur (Spektralkorrektur) der Schwerpunkt. Vier verschiedene Methoden zur Dämpfungskorrektur wurden getestet und verglichen. Die bestimmten Korrekturkoeffizienten unterschieden sich deutlich zwischen den einzelnen Methoden. Jedoch war der Einfluss auf die Absolutwerte halbstündlicher Datensätze gering. Im Gegensatz dazu hatte die Methode deutlichen Einfluss auf die ermittelten Monatssummen der Evapotranspiration. Das zweite Hauptthema beinhaltet einen Vergleich der Wasserbilanz eines Fichten- (Picea abies) mit der eines Buchenbestands (Fagus sylvatica). Beide Bestände befinden sich im Tharandter Wald (Deutschland). Die abiotischen Faktoren sind an beiden Standorten sehr ähnlich. Somit bietet der Vergleich die Möglichkeit Unterschiede in der Wasserbilanz, die durch unterschiedliche Hauptbaumarten verursacht wurden, zu analysieren. Das Ziel was es, die einzelnen Wasserbilanzkomponenten durch eine Kombination der eingangs genanten Messmethoden zu bestimmen und zu vergleichen. Ein Hauptproblem dabei war die Umgehung der unterschiedlichen Messskalen. Deutliche Unterschiede zwischen den beiden Standorten traten nur unter untypischen Wetterbedingungen auf. Unter typischen Bedingungen zeigten die Bestände jedoch ein ähnliches Verhalten. An dieser Stelle erlangten Messunsicherheiten besondere Bedeutung. So demonstrierten die Ergebnisse, dass Unterschiede in der Wasserbilanz beider Standorte durch Messunsicherheiten verwischt wurden. Das dritte Hauptthema behandelt die Wirkung von Unsicherheiten auf Wasserbilanzsimulationen mittels numerischer Modelle. Die Analysen basierten auf Daten von drei Messstationen (Fichten-, Grasland- und Agrarstandort). Es wurde eine Art Monte-Carlo-Simulation eingesetzt, um die Wirkung von Messunsicherheiten zu simulieren. Ferner wurden auch der Einfluss der Modellkomplexität und die Effekte von Unsicherheiten in Referenzdaten auf die Bewertung von Modellergebnissen untersucht. Die Ergebnisse zeigten, dass komplexe Wasserhaushaltsmodelle wie BROOK90 in der Lage sind, das Verhalten und Tendenzen der Wasserbilanz abzubilden. Jedoch wurden zufriedenstellende quantitative Ergebnisse nur unter üblichen Wetterbedingungen erzielt. Unter untypischen Wetterbedingungen (Dürreperioden, Extremniederschläge) wichen die Ergebnisse deutlich vom tatsächlichen (gemessenen) Wert ab. Im Gegensatz zu komplexen Modellen zeigte sich, dass Black Box Modelle (HPTFs) nicht für Wasserhaushaltssimulation an den drei genannten Messstandorten geeignet sind.

Anthropogenic activities are responsible for increases in atmospheric CO₂ and climate change. These increases are partly counterbalanced by natural processes, such as carbon uptake in land surfaces. These processes are themselves subject to climate change, creating a coupled carbon-climate system. I investigated the carbon sink that woody encroachment represents, using a Central Texas savanna as study site, and studied how climatic factors influence this carbon sink. Woody plant encroachment, a worldwide structural change in grassland and savanna ecosystems, alters many ecosystem properties, but the net effect on the carbon balance is uncertain. Woody encroachment represents one of the key uncertainties in the US carbon balance, and demands a more detailed understanding. To come to a process-based understanding of the encroachment effect on carbon dynamics, I analyzed patterns of carbon exchange using eddy-covariance technology. I expected the imbalance between carbon uptake and release processes associated with the encroaching trees specifically, to be responsible for the carbon sink. I also expected that the sink would vary in time, due to strong links between carbon fluxes and soil water in this semi-arid ecosystem. I further studied the ecophysiology of the dominant species, as well as soil respiration processes under different vegetation types, and scaled these findings in space and time. I found that the ecosystem was a significant carbon sink of 405 g C m⁻² yr⁻¹. The encroaching trees increased photosynthesis by 180% and decreased soil respiration by 14%, compared to the grassland, resulting in a strong carbon sink due to the encroachment process. The encroaching process also altered carbon dynamics in relation to climatic drivers. The evergreen species Ashe juniper effectively lengthened the growing season and widened the temperature range over which the ecosystem acts as a carbon sink. The drought resistance of the encroaching trees reduced the sensitivity of this savanna to drought. I conclude that encroachment in Central Texas savannas increased the carbon sink strength by increasing the carbon inputs into the ecosystem. Woody encroachment also reduced the sensitivity to climatic drivers. These two effects constitute a direct effect, as well as a negative feedback to the coupled carbon-climate system. / text

Uncertainty of predicted change in precipitation frequency and intensity motivates the scientific community to better understand, quantify, and model the possible outcome of dryland ecosystems. In pulse dependent ecosystems (i.e. monsoon driven) soil moisture is tightly linked to atmospheric processes. Here, I analyze three overarching questions; Q1) How does soil moisture presence or absence in a shallow or deep layer influence the surface energy budget and planetary boundary layer characteristics?, Q2) What is the role of vegetation on ecosystem albedo in the presence or absence of deep soil moisture?, Q3) Can we develop empirical relationships between soil moisture and the planetary boundary layer height to help evaluate the role of future precipitation changes in land surface atmosphere interactions?. To address these questions I use a conceptual framework based on the presence or absence of soil moisture in a shallow or deep layer. I define these layers by using root profiles and establish soil moisture thresholds for each layer using four years of observations from the Santa Rita Creosote Ameriflux site. Soil moisture drydown curves were used to establish the shallow layer threshold in the shallow layer, while NEE (Net Ecosystem Exchange of carbon dioxide) was used to define the deep soil moisture threshold. Four cases were generated using these thresholds: Case 1, dry shallow layer and dry deep layer; Case 2, wet shallow layer and dry deep layer; Case 3, wet shallow layer and wet deep layer, and Case 4 dry shallow and wet deep layer. Using this framework, I related data from the Ameriflux site SRC (Santa Rita Creosote) from 2008 to 2012 and from atmospheric soundings from the nearby Tucson Airport; conducted field campaigns during 2011 and 2012 to measure albedo from individual bare and canopy patches that were then evaluated in a grid to estimate the influence of deep moisture on albedo via vegetation cover change; and evaluated the potential of using a two-layer bucket model and empirical relationships to evaluate the link between deep soil moisture and the planetary boundary layer height under changing precipitation regime. My results indicate that (1) the presence or absence of water in two layers plays a role in surface energy dynamics, (2) soil moisture presence in the deep layer is linked with decreased ecosystem albedo and planetary boundary layer height, (3) deep moisture sustains vegetation greenness and decreases albedo, and (4) empirical relationships are useful in modeling planetary boundary layer height from dryland ecosystems. Based on these results we argue that deep soil moisture plays an important role in land surface-atmosphere interactions.

Predicting the interactions between climate change and ecosystems remains a core problem in global change research; tropical forest ecosystems are of particular importance because of their disproportionate role in global carbon and water cycling. Amazonia is unique among tropical forest ecosystems, exhibiting a high degree of coupling with its regional hydrometeorology, such that the stability of the entire forest-climate system is dependent on the functioning of its component parts. Belowground ecohydrological interactions between soil moisture environments and the roots which permeate them initiate the water transport pathway to leaf stomata, yet despite the disproportionate role they play in vegetation-atmosphere coupling in Amazonian forest ecosystems, the impacts of climate variability on the belowground environment remain understudied. The research which follows is designed to address critical knowledge gaps in our understanding of root functioning in Amazonian tropical forests as it relates to seasonality and extremes in belowground moisture regime as well as discerning which ecohydrological mechanisms govern ecosystem-level processes of carbon and water flux. A secondary research theme is the evaluation and use of models of ecosystem function as applied to Amazonia - these models are the "knowledge boxes" which build in the ecohydrological hypotheses (some testable than others) deemed to be most important for the forest ecosystems of Amazonia. In what follows, I investigate (i) which mechanisms of water supply (from the soil environment) and water demand (by vegetation) regulate the magnitude and seasonality of evapotranspiration across broad environmental gradients of Amazonia, (ii) how specific hypotheses of root function are or are not corroborated by soil moisture measurements conducted under normal seasonal and experimentally-induced extreme drought conditions, and (iii) the linkage between an extreme drought event with associated impacts on root zone soil moisture, the inferred response of root water uptake, and the observed impacts on ecosystem carbon and water flux in an east central Amazonian forest.