Search

Many photosynthetic species have evolved CO₂-concentrating mechanisms (CCMs) to improve the efficiency of CO₂ assimilation by Rubisco. C3 plants, which include major crops like wheat and rice, lack a CCM and thus, engineering a heterologous CCM into important C3 crops has become a key strategic ambition to enhance yield potential. The green alga Chlamydomonas reinhardtii has the best characterised algal CCM and has become the blueprint for engineering an algal CCM into C3 plants. The Chlamydomonas CCM is based on a Rubisco-dense microcompartment called a pyrenoid, which is supplied with concentrated CO₂ via inorganic carbon transport and carbonic anhydrase activity. Structurally, the pyrenoid is traversed by tubular extensions of the thylakoid membrane, known as the pyrenoid tubules, which are thought to be the site of CO₂ release into the pyrenoid. In this thesis, I assess four bestrophin-like proteins from the Chlamydomonas CCM for their compatibility and function in model C3 plants, to inform efforts to reconstitute a pyrenoid based CCM in C3 plants. In the third chapter, I characterise the first three bestrophin-like proteins (BST1-3), which are thought to be involved in the transport of bicarbonate into the thylakoid lumen. In the current model, bicarbonate then diffuses along the thylakoid lumen and into the pyrenoid tubules. Once in the pyrenoid tubules, bicarbonate is converted to CO₂, which can then diffuse through the tubule membrane into the Rubisco matrix where it is fixed. When I expressed BST1-3 proteins in the C3 plants Arabidopsis thaliana (Arabidopsis) and Nicotiana benthamiana (N. benthamiana), I found they were successfully localised to the thylakoid membrane and assembled into complexes. Further work in Arabidopsis suggested that BST1 and BST2 may function as bicarbonate channels in Arabidopsis but additional validation is required. I also present preliminary evidence that BST1-3 form heterooligomeric channels. Finally, I explored the interaction of BST1-3 with the carbonic anhydrase proteins LCIB and LCIC and found that the previously reported interactions for these proteins are likely mediated by the C-termini of BST1-3. In my fourth chapter, I characterise the fourth bestrophin-like protein, BST4 (also known as Rubisco-binding membrane protein 1), which is hypothesised to function as a linker protein to bring the pyrenoid tubules and the Rubisco matrix together. A heterologous Rubisco condensate in Arabidopsis chloroplasts has already been achieved, and structurally, one of the next steps is to incorporate thylakoid membranes to supply CO₂. I tested whether BST4 could promote the incorporation of thylakoid membrane into this Rubisco condensate in Arabidopsis and found that although BST4 co-localised with the condensate, it was insufficient for the incorporation of thylakoid membrane. Finally, I explored the channel function of BST4 and found that like BST1-3, BST4 localises to the stroma lamellae thylakoid membranes and assembles as a complex. I found that BST4 had adverse effects on plant growth but only when fused to mNeon and co-expressed with the Chlamydomonas Rubisco small subunit or when the C-terminal was mutated or truncated, suggesting BST4 may be subject to C-terminal regulation. However, when BST4 was expressed without a tag in Arabidopsis I found no impact on photosynthetic processes.

Tropical savannas are a dominant biome on Earth characterised by co-dominance of woody plants (trees and shrubs) and grasses that is fundamental to understanding of the global carbon cycle. Tropical savannas display a high level of biogeographic contingency, where both the constituent flora and ecosystem dynamics vary widely among regions, and woody plant composition is dissimilar between continents. To date, there remains a limited understanding of patterns of woody diversity related to climate, fire and drought. Although a new understanding of patterns of diversity and composition will help inform ecosystem models of these regions alongside informing land use policies and highlighting vulnerabilities to climate change. My thesis aims to: 1) quantify patterns of woody diversity in tropical savannas of Africa and Australia in relation to climate and fire; 2) explore use of woody diversity to map the limits of savanna distributions in Africa and Australia; and 3) examine the impacts of drought and fire on community composition in juvenile and adult woody species in a semi-arid savanna in South Africa. To address these aims, I used a combination of data synthesis, experimental field data and statistical modelling. In Chapter one, I introduce the topic and field of research to highlight data gaps related to understanding history, composition and patterns of woody diversity across African and Australian savannas. In Chapter two, using a compilation of 314 plot inventories from Australian and African savannas I quantified the relationship between site level taxonomic richness (alpha diversity) with climate and fire. I found that alpha diversity is higher in African than Australian savannas at any given location. In Africa, patterns of woody diversity have a non-linear relationship with water availability, while in Australia diversity is linearly related to water availability. This chapter is written as a research manuscript to be submitted to the Journal of Biogeography. In Chapter three, I use the same inventory plot data as in Chapter two to quantify woody taxonomic turnover (beta diversity) across gradients of climate and fire. Species turnover is higher over shorter distances in Australia, but decreases for genera and disappears for families, likely due to the dominance of the Eucalyptus and Corymbia genera, and the Myrtaceae family, respectively. This chapter was compiled out of curiosity about complementing the exploration of savanna woody diversity across environmental gradients, but will likely not be published. In Chapter four, using species distribution models and occurrence data for 366 African and 189 Australian savanna woody species, I asked whether savanna distribution can be defined by ensembles of species distributions, and related these ensembles with definitions of savanna based on tree cover and presence of an herbaceous layer. In Australia, mapping savannas using an ensemble of a minimum of 50 species per 0.05 x 0.05° grid cell coincided well with current understanding of savanna limits. However, in Africa where savanna - forest alternative states are more prevalent, the potential to use ensembles of species to map the potential limits of savanna was less clear. This chapter explores the use of the taxonomic data in Chapter two and three to map savanna ecosystems, and is written as a research manuscript to be published in the journal Diversity and Distributions. In Chapter five, using a time series of experimental field data, I quantified the tree demographic responses of juvenile and adult woody species to the 2014-2016 El Nino drought and manipulations of fire frequency in a semi-arid savanna in Kruger National Park, South Africa. Analysing multivariate woody abundance data from 160 sample points over a four year period, it was apparent that juvenile woody diversity increased post-drought on frequently burned and unburned points. However, different fire frequencies favoured different species. Finally, adults suffered substantial drought induced mortality alongside the top-kill from fire. This chapter is written as a research manuscript, which has been accepted in April 2022 as an article in the journal Biotropica. In Chapter six, and in conclusion, I integrate my research findings to suggest that dynamics of woody diversity in savannas of Africa and Australia do not have a universal relationship with environment, likely due to the geography and deep time differences in the development of the biome on each continent. Although, diversity does offer a useful tool to help understand savanna limits. Ecosystem analyses and studies should not treat savannas globally as a single entity given the differences in the woody taxa, unique relationships with environment among continents and the disparity even within species at different life stage to environment.

Climate change induced shifts in precipitation threaten the future carbon balance of Amazon rainforests. Our understanding of water constraints to photosynthesis is largely limited to physiology-climate effects. Less is known about the effects of carbon allocation and trait shifts in response to water availability. Mechanisms linking carbon allocation and trait responses are not well represented in current ecosystem models, causing uncertainty in predicted carbon dynamics under future climates. We ask (i) What drives the coupling between photosynthesis and precipitation is it canopy structure (leaf area index; LAI), leaf traits, or solely a physiology-climate response? (ii) Why do LAI and leaf traits vary with precipitation? (iii) What is the role of water availability and plant traits in driving carbon allocation between leaves and roots? Process based modelling allows the links between photosynthesis, water availability, carbon allocation and traits to be quantified explicitly, exploring interaction space not available to insitu experiments. We calibrated the Soil Plant Atmosphere model (SPA) to eight permanent sample plots across an Amazon mean annual precipitation gradient (1400-2800mm), as part of the Global Ecosystems Monitoring network. The model's representation of local carbon fluxes was evaluated against biometric estimates. We then conducted a series of model experiments to quantify the principal drivers of photosynthesis across the precipitation gradient and explore mechanisms of LAI, leaf trait and carbon allocation responses to water availability. LAI increased with precipitation (R2=0.42, p=0.08), and was the principal driver of differences in photosynthesis across the gradient, accounting for 36% of observed variation. Differences in leaf traits accounted for 20% of variance and physiology-climate interactions accounted for a further 12%. Spatial variance in LAI was underpinned by carbon economics, and best predicted by an optimality approach that maximised net canopy carbon export (R2=0.87, p < 0.001). Across the precipitation gradient, leaf trait strategies shifted from fast to slow as water availability increased (where fast leaf traits are a cohort of high photosynthetic capacitance, high metabolic rate, high nitrogen content, low LMA and short lifespan and vice versa for slow leaf traits). Leaf traits had a determinate effect on LAI optimality, and higher leaf areas at wet plots were supported by longer leaf lifespans rather than an increase in leaf net primary production (NPP). At dry plots, short leaf lifespans, inherent of fast leaf trait cohorts, effected lower LAI. However, fast leaf trait strategies did prove optimal at dry plots, as carbon losses during the dry season were minimised, whilst photosynthesis during the wet season was maximised. Field estimates showed that leaf NPP was highest at dry plots and declined with increasing precipitation, whilst root NPP was highest at wet plots, converse to optimal partitioning theory, which suggests prioritisation of roots under moisture stress and leaves under light limitation. Yet model results show that leaf:root NPP across the precipitation gradient was optimal, and was similarly best predicted by the maximisation of net canopy carbon export (R2=0.60, p=0.02). Optimality was supported by concurrent shifts in leaf and root traits, which together accounted for 63% of variation in optimal leaf:root NPP. Our findings demonstrate that optimality approaches can be used to successfully predict spatial variation in LAI, leaf:root NPP and leaf traits across an Amazon precipitation gradient. Leaf traits fundamentally shaped modelled optimal responses, ultimately determining carbon assimilation. The response of Amazon forests to increased moisture stress is therefore dependent on the current spatial distribution of leaf traits, their plasticity and the likelihood of future shifts in floristic and functional trait composition. Future work should expand on the findings presented by exploring the responses of carbon allocation and traits to water availability over different timescales.

Mangrove forests are important in providing a range of ecosystem services, including food provision to local communities and carbon storage, while being globally restricted to tropical coastlines. The conservation and sustainability of mangrove forests is thus a globally important topic. Mangrove forests in the Niger Delta are known to be under high pressure from urbanisation, development, logging and oil pollution, and invasive species such as nipa palm (Nypa fruticans). These mangrove forests are poorly understood as a result of difficulty of access, social unrest and security restrictions. For example, there is no data on the relationship between disturbance and mangrove structure in the Delta, current area extent and biomass stocks of mangrove forest, its rate of loss, or the rate of nipa palm colonisation in the Niger Delta. The overall objective of this thesis is to utilise a combination of field data and earth observation to resolve these knowledge gaps. This work will estimate area and biomass of mangrove forests in the Niger Delta, and their changes over recent years through disturbance and invasive species. I used an extensive field data collection in 2016-17 to establish 25 geo-referenced 0.25-ha plots across the Niger Delta and collected 567 ground control points. I estimated aboveground biomass (AGB) from a general allometric equation based on stem surveys. Leaf area index (LAI) was recorded using hemispherical photos. I performed and evaluated a land cover classification using a combination of Advanced Land Observatory Satellite Phased Array L-band SAR (ALOS PALSAR), Landsat ETM+ and the Shuttle Radar Topography Mission Digital Elevation Model (SRTM DEM) data. I also compared two supervised classification methods: Maximum Likelihood (ML) and Support Vector machine (SVM) classifiers. I established a relationship between field estimates of AGB and Advanced Land Observatory Satellite (ALOS) L-band radar backscatter. I also estimated the area of nipa palm and mangrove forests in the Niger Delta and generated the first mangrove biomass map for the region, for 2007 and 2017 to obtain change information. Plot estimated mean AGB was 83.7 Mg ha-1 and I found significantly higher plot biomass in close proximity to protected sites and tidal influence, and the lowest in the sites where urbanisation was actively taking place. The mean LAI was 1.45 and there was a significant positive correlation between AGB and LAI (R2= 0.28). Satellite observations of NDVI for the growing season correlated positively with in-situ LAI (R2= 0.63) and AGB (R2= 0.8). Lower stem sizes (5-15cm) accounted for 70% contribution to the total biomass in disturbed plots, while undisturbed plots had a more even contribution of different size classes to AGB. Nipa palm invasion was significantly correlated to plots with larger variations in LAI (i.e. more patchy cover) and proportion of basal area removed within plots. The classification results showed SVM (overall accuracy 99.9 %) performed better than ML (98.7%) across the Niger Delta. I estimated a 2017 mangrove area of 794 561 ha and nipa extent of 11,419 ha. I discovered a 12% decrease in mangrove area and 694 % increase in nipa palm between 2007 and 2017. The highest radar-AGB relationship was from the combination of HH: HV and HV bands (R2= 0.62, p-value < 0.001). Using this relationship, I estimated a mean and total AGB of 90.5 Mg ha-1 and 82 X 106 Mg in 2007; 83.4 Mg ha-1 and 65 X 106 Mg in 2017. Local wood exploitation is removing larger stems (> 15 cm DBH) preferentially from these mangroves and creates an avenue for nipa palm colonisation. I identified opportunities to use remote sensing to estimate biomass, based on the LAI-AGB-NDVI relationship I found, and can serve as a calibration dataset for radar data to provide effective monitoring of mangrove forest degradation. It is clear from these results that remote sensing can be used to map the extent and changes in these land cover types, and thus such mapping efforts should continue for policy targeting and monitoring. I was able to show that mangroves of the Niger Delta are at risk, from rapid clearance as well as from the invasive species nipa palm. I also provide evidence of mangrove cover loss of 11 000 ha yr-1 over a decade, resulting in biomass loss rate of 100 Mg ha-1 yr-1 while mangrove degradation rate of 56 Mg ha-1 yr-1 in the Niger Delta. Assessing carbon stock of mangrove forests in the Niger Delta can create a baseline for regional conservation and regeneration plans. These plans can create opportunities for generating carbon credits under reducing emissions from deforestation and forest degradation (REDD+).

Sudden unpredictable changes in ecosystems are an increasing source of concern because of their inherent unpredictability and the difficulties involved in restoration. Our understanding of the changes that occur across different trophic levels and the form of this change is lacking. This is especially true of large shallow lakes, where characteristics such as fetch and depth are close to theoretical boundary values for hysteretic behaviour. The development of reliable indicators capable of predicting these changes has been the focus of much research in recent years. The success of these early warning indicators (EWIs) has so far been mixed. There remain many unknowns about how they perform under a wide variety of conditions and parameters. Future climate change is predicted to have a wide range of impacts through the interaction of combined pressures, making the understanding of EWIs and the in-lake processes that occur during regime shifts imperative. Loch Leven, Scotland, UK, is a large shallow lake with a history of eutrophication, research and management and as such is an ideal study site to better understand resilience and regime shifts under a range of interacting stressors. The objectives of this research are to: (1) analyse long term data to identify the occurrence of common tipping points within the chemical (water column nutrient concentrations) and biological (macrophytes, phytoplankton, zooplankton) components of the loch, then test these tipping points using five statistical early warning indicators (EWIs) across multiple rolling window sizes; and (2) quantify the changes in lake ecology using a before/after analysis and testing for non-linearity, combined with modelling using the aquatic ecosystem process model PCLake to determine the level of resilience following a regime shift during recovery from eutrophication; (3) using PCLake, examine the sensitivity of Loch Leven to regime shifts in the face of predicted environmental change (e.g. climate change, nutrient pollution). Statistical analysis identified tipping points across all trophic levels included, from physical and chemical variables through to apex predators. The success of EWIs in predicting the tipping points was highly dependent on the number of EWIs used, with window size having a smaller impact. The 45% window size had the highest overall accuracy across all EWIs but only detected 16.5% more tipping points than the window size with the lowest overall accuracy. Differences between individual EWI performance and usage of them as a group was substantial with a 29.7% increase between the two. In both individual and group use of EWIs, false positives (early warning without a tipping point) were more common than true positives (tipping point preceded by EWI), creating significant doubts about their reliability as management tools. Significant change was seen across multiple variables and trophic levels in the before/after analysis following sudden recovery from eutrophication, with most variables also showing evidence of non-linear change. Modelling of responses to nutrient loading for chlorophyll, zooplankton and macrophytes, under states from before and after the shift, indicate hysteresis and thus the presence of feedback mechanisms. The modelling of responses to nutrient loading and predicted climate change in temperature and precipitation demonstrated that increases in temperature and decreases in summer precipitation individually had large impacts on chlorophyll and zooplankton at medium to high phosphorus (P) loads. However, modelling of the combined effects of these changes resulted in the highest lake chlorophyll concentrations of all tested scenarios. At low P loads higher temperatures and increased winter precipitation had the greatest impact on system resilience with a lower Critical Nutrient Load (CNL). The difference between chlorophyll and zooplankton as opposed to macrophytes was in the presence of a lower CNL for the increased winter precipitation-only scenarios which was not seen in the macrophytes. This highlights the potential role of high winter inputs potentially loaded with particulate matter in reducing resilience at lower P loads. This research has highlighted the vulnerability and low resilience of Loch Leven to environmental change. The presence of multiple tipping points and high levels of EWI activity show a high level of flexibility in the system. Coupled with the occurrence of widespread trophic change during a sudden recovery and a small level of hysteresis and high levels of sensitivity to climate change, the low levels of resilience become clear. The impact of lake-specific characteristics such as moderate depth, large fetch and a heterogeneous bed morphology is particularly evident in the limitations on macrophyte cover and the reliance on zooplankton to determine the hysteresis offset (amount of phosphorus (P) loading between the two CNL). The presence of these characteristics can be used to identify other lakes vulnerable to change. Improving the predictive capabilities of resilience indicators such as EWIs, and better understanding of the ecological changes that occur during non-linear change in response to recovery and climate change, can help target relevant ecosystem components for preventative management. These actions may become necessary under even the most conservative estimates of environmental change.

The tundra biome has undergone dramatic vegetation shifts in recent decades, which have been partly attributed to climate warming. Shrub species in particular are expanding widely throughout the Pan-Arctic region, and are involved in complex vegetation-atmosphere interactions that have important implications for the global energy balance and carbon budget. However, projections of vegetation change and associated feedbacks are complicated by the high variability in the sensitivity of shrub growth to temperature among sites and species. A mechanistic understanding of the individual-to-regional controls of climate sensitivity is therefore needed to accurately predict future vegetation change at the biome scale. This thesis quantifies the influence of environmental and ecological factors, and especially of plant-plant interactions, on the growth response of Arctic shrub communities to climate change. Climate change in the Arctic has resulted in warmer, but also longer growing seasons in many locations due to earlier snowmelt. These two factors are often treated as one single control of plant growth, but with scarce records of green-up and senescence dates for the Arctic, few studies have measured the sensitivity of shrub growth to changes in growing season length. Using radial growth time series from over 300 shrubs collected at four sites of contrasting climatic regimes and greening trajectories in Northern Canada, I measured the sensitivity of shrub growth to summer temperature and satellite-derived growing season length. I found that growing season length and summer temperature were decoupled within sites and had inconsistent effects on growth across the four sites. My findings indicate that longer and warmer growing seasons do not necessarily act as combined drivers of vegetation change across the biome. My research also demonstrated that growth at the root collar of shrubs is more climate sensitive than stem growth, possibly indicating differential internal resource allocation strategies, and highlighting the importance of standardised protocols when comparing dendroecological data across multiple sites. Individual and species traits are thought to play an important role in the response of tundra vegetation to climate change. Taller shrub species have been shown to be more climate-sensitive than dwarf shrubs, but whether this relationship holds at the individual level is unknown. I tested whether plant size, as a proxy for competitive ability, explained variation in the climate sensitivity of shrub growth using 1085 individual size and growth-ring records from 16 species at 18 sites across the tundra biome. I did not find evidence that taller shrubs were more climate sensitive, and found that height became a progressively poorer predictor of other growth dimensions at higher latitudes. This suggests that predictions of functional and structural change based on allometric equations from boreal or sub-Arctic populations may not be valid for the tundra biome as a whole. Plant-plant interactions are a strong driver of community dynamics. With increasing shrub densities in the circumpolar region, competition could have an increasingly important influence on shrub growth, potentially limiting climate-driven expansion. I found that competition with trees might slow down shrub expansion in the boreal forest biome, as the climate sensitivity of shrub growth was much lower in a boreal forest in southwest Yukon compared to shrubs growing in the alpine tundra in the same region. However, my findings did not indicate a strong control of shrub-shrub competition on growth. A canopy removal experiment did not reveal any difference in the growth rate of shrubs having experienced a decrease in aboveground competition compared to shrubs growing in intact shrub patches. Additionally, shrubs experiencing more competition were generally as climate sensitive as those with fewer or more distant neighbours, as I demonstrated through spatial analysis at four sites across the Canadian Arctic. However, their spatial arrangement, with positive size-distance relationships between pairs of neighbours, suggested that competition does play a role in the life history of these shrubs, especially at more productive sites. Finally, I found evidence of physical and chemical interference of ground vegetation on the germination of deciduous shrub seeds, indicating that interactions with other plant functional groups may control rates of shrub expansion. Shrub expansion at the plot to landscape scale has been heavily documented over multiple decades through several lines of evidence including long-term monitoring, remote sensing, and experimental studies. The increase in shrub biomass in the tundra has high certainty both in detection and in attribution to climate warming. However, my thesis highlights the complexity and variability of growth responses when using radial growth as an indicator of climate sensitivity. I detected this variability at multiple scales, from plant parts within an individual showing inconsistent climatic signals, to site-scale sensitivity responding to different facets of global change. I did not find strong or consistent influences of biotic and abiotic controls on the growth responses of tundra shrubs; however, these relationships may change over time as shrub densities continue to increase and exacerbate resource limitations. With 80% of tundra biomass potentially located below ground, understanding whole-plant and community-level responses to climate will be critical to improve projections of tundra plant community responses to global change. Understanding the different drivers of primary and secondary growth will be key to using estimates of climate sensitivity derived from growth-ring records to project biomass change and associated feedbacks across the tundra biome.

The tundra is currently warming twice as rapidly as the rest of planet Earth, which is thought to be leading to widespread vegetation change. Understanding the drivers, patterns, and impacts of vegetation change will be critical to predicting the future state of tundra ecosystems and estimating potential feedbacks to the global climate system. In this thesis, I used plant traits - the characteristics of individuals and species - to investigate the fundamental structure of tundra plant communities and to link vegetation change to decomposition across the tundra biome. Plant traits are increasingly used to predict how communities will respond to environmental change. However, existing global trait relationships have largely been formulated using data from tropical and temperature environments. It is thus unknown whether these trait relationships extend to the cold extremes of the tundra biome. Furthermore, it is unclear whether approaches that simplify trait variation, such as the categorization of species into functional groups, capture variation across multiple traits. Using the Tundra Trait Team database - the largest tundra trait database ever compiled - I found that tundra plants revealed remarkable consistency in the range of resource acquisition traits, but not size traits, compared to global trait distributions, and that global trait relationships were maintained in the tundra biome. However, trait variation was largely expressed at the level of individual species, and thus the use of functional groups to describe trait variation may obscure important patterns and mechanisms of vegetation change. Secondly, plant traits are related to several key ecosystem functions, and thus offer an approach to predicting the impacts of vegetation change. Notably, understanding the links between vegetation change and decomposition is a critical research priority as high latitude ecosystems contain more than 50% of global soil carbon, and have historically formed a long-term carbon sink due to low decomposition rates and frozen soils. However, it is unclear to what extent vegetation change, and thus changes to the quality and quantity of litter inputs, drives decomposition compared to environmental controls. I used two common substrates (tea), buried at 248 sites, to quantify the relative importance of temperature, moisture and litter quality on litter decomposition across the tundra biome. I found strong linear relationships between decomposition, soil temperature and soil moisture, but found that litter quality had the greatest effect on decomposition, outweighing the effects of environment across the tundra biome. Finally, I investigated whether tundra plant communities are undergoing directional shifts in litter quality as a result of climate warming. Given the importance of litter quality for decomposition, a shift towards more or less decomposable plant litter could act as a feedback to climate change by altering decomposition rates and litter carbon storage. I combined a litter decomposition experiment with tundra plant trait data and three decades of biome-wide vegetation monitoring to quantify change in community decomposability over space, over time and with warming. I found that community decomposability increased with temperature and soil moisture over biogeographic gradients. However, I found no significant change in decomposability over time, primarily due to low species turnover, which drives the majority of trait differences among sites. Together, my thesis findings indicate that the incorporation of plant trait data into ecological analyses can improve our understanding of tundra vegetation change. Firstly, trait-based approaches capture variation in plant responses to environmental change, and enable prediction of vegetation change and ecosystem function at large scales and under future growing conditions. Secondly, my findings offer insight into the potential direction, rate and magnitude of vegetation change, indicating that despite rapid shifts in some traits, the majority of community-level trait change will be dependent upon the slower processes of migration and species turnover. Finally, my findings demonstrate that the impact of warming on both tundra vegetation change and ecosystem processes will be strongly mediated by soil moisture and trait differences among vegetation communities. Overall, my thesis demonstrates that the use of plant traits can improve climate change predictions for the tundra biome, and informs the fundamental rules that determine plant community structure and change at the global scale.

Nitrous oxide (N2O) is a powerful greenhouse gas and a major contributor to ozone layer depletion. The application of nitrogenous fertilisers to agricultural soils is a major source of N2O on a global scale. Arable soils receive significant rates of synthetic nitrogen (N) and thus have a considerable N2O footprint. The reduction of the N2O footprint of agricultural systems is a key target for those countries that seek to reduce their contribution to climate change and achieve a more sustainable agriculture. These twin targets are part of Scotland's agro-environmental policy. Because soil N2O emissions vary significantly both temporally and spatially, measuring N2O emissions across wide agricultural areas is impractical. However, the quantification of the N2O footprint of important agricultural regions is very valuable to scientists, farmers and policymakers alike. In this context, agro-ecosystem biogeochemistry models are scientific tools, which are developed using in-depth knowledge on the underlying processes, and are used to quantify N2O emissions across spatial and temporal scales. In Scotland, arable agriculture is concentrated at the Eastern part of the country where wheat, barley and oilseed rape are the most widely cultivated crops. The main aim of this study was to quantify the amount of N2O that is emitted from arable soils due to the cultivation of these three crops in Eastern Scotland by using the Landscape-DNDC model. Landscape-DNDC is a mechanistic biogeochemistry model that describes the flows of energy, water and nutrients in agricultural ecosystems. As part of the study, the parametric sensitivities of key model outputs have been quantified using well-established sensitivity analysis methods, which were tailored in order to consider the particularities of N cycling in arable soils. Driven by the fact that the existence of spatiotemporal uncertainties around field-measured soil N2O data complicates the evaluation of model performance, a novel model evaluation algorithm has been developed and was used to assess the model's predictive accuracy. By combining the knowledge of the model's parametric sensitivity with the abilities of the evaluation algorithm, nine key parameters of Landscape-DNDC were calibrated to UK edaphoclimatic conditions (using the Metropolis-Hastings Bayesian calibration algorithm). Model calibration led to improved prediction of field-measured soil N2O emissions at a set of sites. The model was then coupled to geographically explicit data on climate, soil N2O and crop management and used to simulate N2O emissions from the arable soils of Eastern Scotland. The results show that, on average, 0.59 % of the applied fertiliser N (kg N ha-1) was lost to the atmosphere as N2O. This factor is much lower than the generic N2O emission factor (EF) of 1% and closer to the UK cropland-specific N2O EF (i.e. 0.79%). The predicted annual N2O was the combined result of different drivers (i.e. fertiliser rate, soil and climate variables) but the geographic distribution of the estimated N2O EFs revealed some hotspots of high N2O EF (larger than 1%). Interestingly, these hotspots were caused by the cultivation of winter oilseed rape on soils with high bulk density and clay content. The comparison of the simulated yields per hectare with respective measured data and of the simulated nitrate (NO-3 ) leaching and crop N uptake factors with respective literature-based values showed that the prediction of soil N2O was not made at the expense of realistic prediction of other important aspects of agro-ecosystem biogeochemistry. Also, the study found that the simulated N2O is almost twice as sensitive to soil input uncertainty as the simulated NO-3 is, while, crop N uptake is rather insensitive to this source of uncertainty. Finally, the study shows that the uncertainty around the nine calibrated model parameters affects the prediction of NO-3 leaching strongly but its role in regards to the simulation of N2O emissions is small.

The terrestrial CO2 exchange in the Arctic plays an important role in the global carbon (C) cycle. The Arctic ecosystems, containing a large amount of organic carbon (C), are experiencing ongoing warming in recent decades, which is affecting the C cycling and the feedback interactions between its different components. To improve our understanding of the atmosphere-ecosystem interactions, the Greenland Ecosystem Monitoring (GEM) program measures ecosystem CO2 exchange and links it to biogeochemical processes. However, this task remains challenging in northern latitudes due to an insufficient number of measurement sites, particularly covering full annual cycles, but also the frequent gaps in data affected by extreme conditions and remoteness. Combining ecosystem models and field observations we are able to study the underlying processes of Arctic CO2 exchange in changing environments. The overall aim of the research is to use data-model approaches to analyse the patterns of C exchange and their links to biological processes in Arctic ecosystems, studied in detail both from a measurement and a modelling perspective, but also from a local to a pan-arctic scale. In Paper I we found a compensatory response of photosynthesis (GPP) and ecosystem respiration (Reco), both highly sensitive to the meteorological drivers (i.e. temperatures and radiation) in Kobbefjord, West Greenland tundra. This tight relationship led to a relatively insensitive net ecosystem exchange (NEE) to the meteorology, despite the large variability in temperature and precipitations across growing seasons. This tundra ecosystem acted as a consistent sink of C (-30 g C m-2), except in 2011 (41 g C m-2), which was associated with a major pest outbreak. In Paper II we estimated this decrease of C sink strength of 118-144 g C m-2 in the anomalous year (2011), corresponding to 1210-1470 tonnes C at the Kobbefjord catchment scale. We concluded that the meteorological sensitivity of photosynthesis and respiration were similar, and hence compensatory, but we could not explain the causes. Therefore, in Paper III we used a calibrated and validated version of the Soil-Plant-Atmosphere model to explore full annual C cycles and detail the coupling between GPP and Reco. From this study we found two key results. First, similar metrological buffering to growing season reduced the full annual C sink strength by 60%. Second, plant traits control the compensatory effect observed (and estimated) between gross primary production and ecosystem respiration. Because a site-specific location is not representative of the entire Arctic, we further evaluated the pan-Arctic terrestrial C cycling using the CARDAMOM data assimilation system in Paper IV. Our estimates of C fluxes, pools and transit times are in good agreement with different sources of assimilated and independent data, both at pan-Arctic and local scale. Our benchmarking analysis with extensively used Global Vegetation Models (GVM) highlights that GVM modellers need to focus on the vegetation C dynamics, but also the respiratory losses, to improve our understanding of internal C cycle dynamics in the Arctic. Data-model approaches generate novel outputs, allowing us to explore C cycling mechanisms and controls that otherwise would not have been possible to address individually. Also, discrepancies between data and models can provide information about knowledge gaps and ecological indicators not previously detected from field observations, emphasizing the unique synergy that models and data are capable of bringing together.

The representation of high-latitude surface processes and quantifying surface-climate feedbacks are some of the most serious shortcomings of present day Arctic land surface modelling. The energy balance of seasonally snow-covered sparse deciduous forests at high latitudes is poorly understood and inaccurately represented within hydrological and climate models. Snow cover plays an important role in wintertime fluxes of energy, water and carbon, controlling the length of the active growing season and hence the overall carbon balance of Arctic ecosystems. Snow cover is non-uniform and spatially variable, as wind redistributes snow from areas of exposed open tundra to sheltered areas within the forest, where a deeper snowpack develops. Low solar zenith angles, coupled with sparse deciduous leafless trees, cast shadows across the snow surface. The spatial distribution of canopy gaps determines the timing of direct radiation which penetrates down through the canopy to the snow surface. The forest canopy also excludes incoming longwave radiation and yet also emits longwave radiation to the snow surface. Consequently the forest canopy plays a key role in the radiation balance of sparse forests. To improve our knowledge of these complex processes, meteorological and field observations were taken in an area of highly heterogeneous birch Betula pubescens ssp. czerepanovii forest in Abisko, Sweden during the spring of 2008 and 2009. Detailed measurements of short and longwave radiation above and below the canopy, hemispherical photographs, tree temperatures and snow surveys were conducted to quantify the radiation balance of the sparse deciduous forest. An array of below canopy pyranometers found the mean canopy transmissivity to be 74 % in 2008 and 76 % in 2009. Hemispherical photographs taken at the pyranometer locations analysed with Gap Light Analyzer (GLA) showed reasonable agreement with a mean canopy transmissivity of 75 % in 2008 and 74 % in 2009. The canopy transmissivity was found to be independent of the diffuse fraction of radiation as the canopy is very sparse. A series of survey grids and transects were established to scale up from the below canopy pyranometers to the landscape scale. Hemispherical photographs analysed with GLA showed the sparse forest canopy had a mean transmissivity of 78 % and a mean LAI of 0.25, whereas the open tundra had a mean transmissivity of 97 % and a mean LAI of < 0.01. Snow surveys showed the sparse forest snow depth to vary between 0.34 and 0.55 m, whereas the snow depth in the open tundra varied between 0.12 and 0.18 m. Observations of canopy temperatures showed a strong influence of incident shortwave radiation warming the tree branches to temperatures up to 15 °C warmer than ambient air temperature on the south facing sides of the trees, and up to 6 °C on the north facing sides of the trees. To reproduce the observed radiation balance, two canopy models (Homogenous and Clumped) were developed. The Homogeneous canopy model assumes a single tree tile with a uniform sparse canopy. The Clumped canopy model assumes a tree and a grass tile, where the tree tile is permanently in shade from the canopy and the grass tile receives all the incoming radiation. These canopy models identified the need for a parameter that accounts for the spatial and temporal variation of the shaded gaps within the sparse forest. JULES (Joint UK Land Environment Simulator) is the community land surface model used in the UK Hadley Centre GCM suite. Modifications of the land-surface interactions were included in JULES to represent the shaded gaps within the sparse deciduous forest. New parameterisations were developed for the time-varying sunlit fractions of the gap (flit), the sky-view fraction (fv), and the longwave radiation emitted from the canopy (LWtree). These model developments were informed by field observations of the forest canopy and evaluated against the below canopy short and longwave radiation observed data sets. The JULES Shaded gap model output showed a strong positive relationship with the observations of below canopy shortwave and longwave radiation. The JULES Shaded gap model improves the ratio of observed to modelled short and longwave radiation on sunny days compared to the JULES model. The JULES Shaded gap model reduces the time to snow melt by 2 to 4 days compared to the JULES model, making the model output more aligned with in-situ observational data. This shortening of the modelled snow-season directly impacts on the simulated carbon and water balance regionally and has wider relevance at the pan-Arctic scale. When JULES Shaded Gap was evaluated on the global scale, it improved the modelled snowmass across large areas of sparse forest in northern Canada, Scandinavia and Northern Russia with respect to GlobSnow. The performance of the land surface-snow-vegetation interactions of JULES was improved by using the Shaded gap to model the radiation balance of sparse forests in climate-sensitive Arctic regions. Furthermore these observational data can be used to develop and evaluate high latitude land-surface processes and biogeochemical feedbacks in other earth system models.

Savannah ecosystems play a prominent role in the global carbon (C) cycle, yet fluxes are poorly quantified, and the key processes regulating vegetation dynamics are uncertain. Insight is particularly deficient in southern Africa’s miombo woodlands, a woody savannah that is home to over 100 million people. This biome is heavily disturbed, with widespread deforestation and degradation associated with agriculture, charcoal and timber extraction, and frequent fires from anthropogenic sources. In this thesis I combine plot inventory data with remote sensing and modelling techniques to improve our understanding of the miombo woodland C cycle. Using a network of forest inventory plots, I characterise floristic and functional diversity in a savannah-forest mosaic in southeastern Tanzania. Divergent vegetation structures are associated with variation in fire frequency, water supply, and soil chemo-physical properties. Corresponding differences are noted in fire resilience, water-use, and nutrient acquisition plant functional traits, suggesting that multiple interrelated environmental filters act to assemble heterogeneous tree communities. Re-inventory of forest plots was used to quantify key aspects of the woody C cycle. Tree growth rates are slow, calling for careful management of woodland resources, and significantly reduced where stems were damaged. Stem mortality is rare, though elevated in the smallest trees and where damage was recorded. Contemporary strategies to incentivise the conservation of miombo woodland ecosystems, such as the REDD+ programme of the United Nations, advocate payments for sustaining ecosystem services such as C sequestration. I report on a pilot REDD+ project aiming to reduce woodland degradation from frequent high intensity fires in southeastern Tanzania. Model simulations suggest that woody biomass is being gradually lost from the region, and that setting early season fires has the potential to reverse this trend. Realising substantial changes in C storage requires a demanding reduction to late fire frequency, and uncertainty in model predictions remains high. I quantify the C cycle of southern African woodlands by combining observational data with a diagnostic C cycle model under a model-data fusion framework. Model outputs show substantial variation in primary production, C allocation patterns, and foliar and canopy traits, which are associated with differences in woody cover, fire, and precipitation properties. C cycle dynamics correspond poorly to conventional land cover maps, indicating they may be unsuited to upscaling measurements and models of the terrestrial C cycle.

Natural tropical deciduous forest (TDF) is considered with a medium to small height (< 15 m). Particularly, in Mexico TDF shows a remnant of 36.2% of primary forest driving changes in the structure and species composition. This vegetation in Mexico is mainly transformed into grassland for cattle raising, and agriculture, primarily for self-consumption. More information about the ecology and the social pressures on this vegetation can be seen in Chapter I. The general methods, including sampling allocation and collection, characteristics of the study site, as well the procedure of the research proposal is presented in Chapter II. The main aim of this thesis is to improve the accuracy of predictions of net carbon emissions and the spatial distribution of AGB in the Tropical Deciduous Forest of Mexico. To address this aim, it is important to take into consideration the forest structure, spatial patterns and processes in a natural forest in a multi-scale analysis; also, it is necessary to characterize the spatial socio-economic drivers that influence current AGB losses. With the understanding of such elements, it is possible to reconstruct the potential carbon stocks and estimate the allocation of net carbon emissions due to deforestation and forest degradation. This study shows that it is possible to count net carbon emissions caused by deforestation and forest degradation at a landscape scale. To come to such estimates, it was necessary to reduce the different sources of uncertainty. Chapter III explores different elements that drive the AGB allocation in a mature forest. The AGB in the mature forest was considered as the potential AGB that the forest could get assuming that it has reached its steady state. Different field sampling strategies and allometric equations were evaluated to account for uncertainty in the AGB estimations. The results showed that small sampling design (300-400 m2) and large-sized plots (4 ha) produce the same tree distribution for trees: ≥30 cm in DBH as well as in AGB. These results contradict what has been reported for others (Chave et al., 2004 and 2005) when they refer to the general definition of tropical forest. However, those other studies referred to forests with a much higher precipitation and which can be classified as tropical rain (perennial) forest (Chave et al., 2004). In the tropical deciduous forest, the kind considered in this study, AGB tends to be allocated in small-sized trees. Diverse biophysical characteristics that may drive AGB allocation were considered over different spatial scales. Water stress was the main driver for AGB density at different spatial scales. Nutrients showed little significance to explain AGB as other studies have suggested in secondary forests and/or chronosequences. With this understanding, Chapter IV shows the use of different multi-variable models. Parsimonious models were the result of the variables selection and sensitivity test. Most of the methodologies showed a better performance to explain AGB allocation than a null-model. However, when they were contrasted with independent observations over different spatial resolutions, it was possible to conclude that only GLM was capable of reproducing the spatial patterns, and its estimations were close to observations. Nevertheless, some observations with very large AGB densities were underestimated by the model. This underestimation was related to the presence of few very large-sized trees. These two chapters depict the possibility of accounting for the potential AGB, and the uncertainty, namely whether the landscape could reach it with the absence of human disturbance. Once the potential AGB map was built and validated, it was transformed to carbon stock, using a local carbon concentration estimate. This potential carbon stock map was contrasted to the different available maps of current carbon stocks. Consequently, it was possible to estimate net carbon emissions due to deforestation and forest degradation (Chapter V), suggesting that the general models tend to agree in the total carbon loss. However, there are some spatial discrepancies in the magnitudes of change. Main differences between maps can be reduced by diverse socio-ecological constraints that dominate the landscape. This is important because it may be possible to make future adjustments that would reduce variability, enabling more accurate AGB estimations. However, to individually account for deforestation and forest degradation, more detailed sources of local information are necessary, such as socio-economic variables. Therefore models with a bottom-up perspective would lead to a better understanding and representation of the landscape. Finally, the growing rural population will have larger demands for wood and food, so while remote or protected areas may have the potential for storing high AGB, forest close to settlements and access routes are likely to continue being disturbed, unless affordable alternatives are available for the sustainable use of the forest. In conclusion, the estimation of spatial heterogeneity of AGB in the landscape is of great importance when measuring carbon stocks and ecological dynamics. Various elements influence the AGB allocation in the mature forest. Among all of them, water availability played the most decisive part of various spatial scales. My models support the hypothesis that water availability plays the major role in explaining AGB in Mexico on a local, sub-regional and landscape scale. Model selection produced contrasting AGB estimates and patterns. Moreover, the results of this study tell us that there is not a clear consensus among various current AGB maps. However, they also show that with a multi-model comparison it is possible to identify carbon emissions drivers and calculate total carbon emissions due to forest disturbances. Socio-economic variables played the major role in explaining AGB losses. Therefore, future studies should look into a bottom-up approach for a better understanding and representation of current AGB.

Wind is the main abiotic cause of disturbance to forests in large parts of the world, particularly at temperate and boreal latitudes. In the past few decades the consequences of large wind-induced losses have been experienced at various levels, from small forest owners to large-scale, whole-society level. This is particularly relevant for areas, such as Europe, where forests are intensively managed, and the assets exposed to wind hazard are substantial. To better manage forests and commercial tree plantations to reduce the risk of wind damage, process-based, semi-mechanistic mathematical models such as ForestGALES are used. This model has been parameterised and evaluated for numerous conifer species, which constitute the major plantation types in temperate and boreal biomes. However, the geographical extent and economic importance of fast-growing broadleaved species, such as those of the Eucalyptus genus, and the lack of detailed historical data on wind damage to these species, require that tools for the estimation of the risk of wind damage to these species are developed and evaluated. This is particularly relevant in light of the projected increases of surface temperature due to climate change, and of the frequency and severity of extreme windstorms, that are expected as a consequence of climate change. Fieldwork was conducted in a semi-natural Eucalyptus globulus (Labill.) forest in the Asturias region in Northern Spain to acquire data for the parameterisation of ForestGALES for E. globulus, using a tree-pulling experiment. The behaviour of the parameterisation was investigated for different stocking densities to evaluate whether the effects of tree height, stocking density, and presence of a fresh upwind gap are consistent with the literature. This parameterisation was then used to compare the vulnerability to wind damage between E. globulus and Pinus pinaster (Ait.), the predominant plantation species in the Aquitaine region of SW France where extensive damage was experienced from storms Martin (1999) and Klaus (2009). The effects of rooting depth (2x), growth rate (2x), presence/absence of a recently created windward gap, and of the predominant wind climate in Aquitaine were investigated in this comparison. In order to aid forest managers with optimal resource allocation for practical applications of ForestGALES, and to provide forest modellers with invaluable insights for the development of robust wind damage risk models, ForestGALES was subjected to a sensitivity analysis. A generalisation of the variance-based method of Sobol’ for the case of correlated variables was used to investigate the sensitivity of the outputs of ForestGALES (the critical wind speeds for stem breakage and uprooting, and the associated probabilities of damage) to variation in its input variables. Almost all the E. globulus trees pulled in Asturias failed by overturning rather than breakage, which allowed for good confidence in the calculations of the overturning moments required for the empirical component of ForestGALES. Resistance to overturning was not significantly influenced by the presence of a tap-root. Modelling the shape of the tree crowns with an ellipsoid provided a good approximation of the geometry of the canopy, but required additional fieldwork as crown width in the four cardinal directions had to be estimated visually prior to the tests. The scarcity of detailed published data on wind damage to E. globulus made evaluating the parameterisation particularly challenging. This impediment was obviated by investigating the behaviour of the parameterisation with regards to the well-known effects of tree height, stocking density, and presence of a fresh upwind gap. The simulations showed that the parameterisation behaved as expected, with vulnerability of E. globulus stands increasing with tree height, stocking density, and the presence of a gap. High initial planting densities, an early thinning, and a final harvesting before the trees have reached a height of 20 – 25m are recommended to reduce the risk of wind damage to E. globulus. The comparison with P. pinaster showed that E. globulus trees are particularly susceptible to the presence of a recently created windward gap. Therefore, harvesting at neighbouring sites should be minimised, and preferentially performed when the neighbouring stands are still at a young age to take advantage of the fast growth rates of E. globulus. These practices would ensure that in case of wind damage any losses are recovered in a short time. These procedures can reduce the cumulative risk through the rotation, while maintaining competitive yields. The ForestGALES simulations have also highlighted that the silvicultural practices currently in place in Aquitaine expose P. pinaster trees to high levels of cumulative risk (> 20%). The sensitivity analysis of ForestGALES has highlighted the strengths of the model and the areas that require substantial improvement. The results of the analysis show that ForestGALES is able to simulate very effectively the dynamics of wind damage to forest stands, as the model architecture reflects the significant influences of tree height, stocking density, dbh, and size of an upwind gap, on the calculations of the critical wind speeds of damage. Similarly, in ForestGALES the wind climate of a site is the main driver of variation of the probabilities of damage, as it is for real forests affected by extreme storms. Conversely, when the windiness of a site is moderate, ForestGALES accounts for the larger role of tree and stand variables. The sensitivity analysis has shown that ForestGALES is particularly efficient at simulating not only the effect of the size of windward gaps on the vulnerability of a stand, but also at differentiating between recently formed stand edges and edges that have been in place since the establishment of a stand. Therefore, for practical applications of the model, tree height, dbh, stocking density, the size and nature of an upwind gap, and the local wind climate, are the variables that need to be known with a high accuracy in order to maximally reduce the uncertainty of the model predictions. The section of the model that requires further attention and research is the one dedicated to the calculation of the trees’ resistance to overturning. The sensitivity analysis has shown that rooting depth and soil type, the model input variables on which the empirical component of ForestGALES that describes the resistance to overturning is based, contribute only marginally to the variation in the outputs. This finding unequivocally identifies that efforts for future research should be aimed at studying the mechanics of root-soil interactions with regards to tree stability. The results of the sensitivity analysis have also shown that the variance-based method used in this research project is equally sensitive to the accurate description of the probability distribution functions of the scrutinised variables, as it is to their correlation structure.

Cropland management practices have traditionally focused on maximising the production of food, feed and fibre. However, croplands also provide valuable regulating ecosystem services, including carbon (C) storage in soil and biomass. Consequently, management impacts the extents to which croplands act as sources or sinks of atmospheric carbon dioxide (CO2). And so, reliable information on cropland ecosystem C fluxes and yields are essential for policy-makers concerned with climate change mitigation and food security. Eddy-covariance (EC) flux towers can provide observations of net ecosystem exchanges (NEE) of CO2 within croplands, however the tower sites are temporally and spatially sparse. Process-based crop models simulate the key biophysical mechanisms within cropland ecosystems, including the management impacts, crop cultivar, soil and climate on crop C dynamics. The models are therefore a powerful tool for diagnosing and forecasting C fluxes and yield. However, crop model spatial upscaling is often limited by input data (including meteorological drivers and management), parameter uncertainty and model complexity. Earth observation (EO) sensors can provide regular estimates of crop condition over large extents. Therefore, EO data can be used within data assimilation (DA) schemes to parameterise and constrain models. Research presented in this thesis explores the key challenges associated with crop model upscaling. First, fine-scale (20-50 m) EO-derived data, from optical and radar sensors, is assimilated into the Soil-Plant-Atmosphere crop (SPAc) model. Assimilating all EO data enhanced the simulation of daily C exchanges at multiple European crop sites. However, the individually assimilation of radar EO data (as opposed to combined with optical data) resulted in larger improvements in the C fluxes simulation. Second, the impacts of reduced model complexity and driver resolution on crop photosynthesis estimates are investigated. The simplified Aggregated Canopy Model (ACM) – estimating daily photosynthesis using coarse-scale (daily) drivers – was calibrated using the detailed SPAc model, which simulates leaf to canopy processes at half-hourly time-steps. The calibrated ACM photosynthesis had a high agreement with SPAc and local EC estimates. Third, a model-data fusion framework was evaluated for multi-annual and regional-scale estimation of UK wheat yields. Aggregated model yield estimates were negatively biased when compared to official statistics. Coarse-scale (1 km) EO data was also used to constrain the model simulation of canopy development, which was successful in reducing the biases in the yield estimates. And fourth, EO spatial and temporal resolution requirements for crop growth monitoring at UK field-scales was investigated. Errors due to spatial resolution are quantified by sampling aggregated fine scale EO data on a per-field basis; whereas temporal resolution error analysis involved re-sampling model estimates to mimic the observational frequencies of current EO sensors and likely cloud cover. A minimum EO spatial resolution of around 165 m is required to resolve the field-scale detail. Monitoring crop growth using EO sensors with a 26-day temporal resolution results in a mean error of 5%; however, accounting for likely cloud cover increases this error to 63%.

Climate change is arguably the most important global societal challenge. Developing ‘low-carbon societies’, i.e. reducing greenhouse gas (GHG) emissions and adapting to a changing climate, is becoming a policy goal across the globe. Agriculture plays an important role in this transformation. The sector is highly vulnerable to climate variability, and is a significant source of emissions. At the same time, it has potential for reducing GHG emissions and also provides opportunity for carbon sequestration in soils and crop biomass. Policy support for mitigating GHG emissions is being informed by scientific evidence on the effectiveness and costs of mitigation opportunities. This information is frequently depicted in marginal abatement cost curves (MACCs), an assessment tool which can help to visualise the hierarchy of technical measures and their cumulative level of abatement. Similarly to other assessment tools, MACCs’ suitability to provide information has certain limitations. Furthermore, different derivations of MACCs are appropriate to answer different questions. In order to draw both informative and reliable conclusions for policy decisions, the characteristics of the MACCs and the resulting limitations have to be presented clearly. This dissertation seeks to answer the general question whether the agricultural MACCs can be improved so that they provide more comprehensive and tailored information to policy makers. In particular five limitations of the MACCs are discussed: the lack of representation of wider effects, the issue of cost-effectiveness of policy instruments and the inclusion of transaction costs, the uncertainty in the MACCs, the boundaries and the heterogeneity of the analysis. Theoretical frameworks are developed and case study examples are provided for these limitations, and the frameworks are assessed in terms whether they achieve the goal of providing more comprehensive information to policy makers than a conventional MACC. Furthermore, the dissertation summarises the available methodologies and applications in agriculture to enhance the MACCs and provides guidelines for researchers and policy makers about the choice of methods and the communication of the results in order to improve the use of MACCs in the policy process.

Forestry in the UK is dominated by coniferous plantations. Sitka spruce (Picea sitchensis) and Scots pine (Pinus sylvestris) are the most prevalent species and are mostly grown in single age mono-culture stands. Forest strategy for Scotland, England, and Wales all include efforts to achieve further afforestation. The aim of this afforestation is to provide a multi-functional forest with a broad range of benefits. Due to the time scale involved in forestry, accurate forecasts of stand productivity (along with clearly defined uncertainties) are essential to forest managers. These can be provided by a range of approaches to modelling forest growth. In this project model comparison, Bayesian calibration, and data assimilation methods were all used to attempt to improve forecasts and understanding of uncertainty therein of the two most important conifers in UK forestry. Three different forest growth models were compared in simulating growth of Scots pine. A yield table approach, the process-based 3PGN model, and a Stand Level Dynamic Growth (SLeDG) model were used. Predictions were compared graphically over the typical productivity range for Scots pine in the UK. Strengths and weaknesses of each model were considered. All three produced similar growth trajectories. The greatest difference between models was in volume and biomass in unthinned stands where the yield table predicted a much larger range compared to the other two models. Future advances in data availability and computing power should allow for greater use of process-based models, but in the interim more flexible dynamic growth models may be more useful than static yield tables for providing predictions which extend to non-standard management prescriptions and estimates of early growth and yield. A Bayesian calibration of the SLeDG model was carried out for both Sitka spruce and Scots pine in the UK for the first time. Bayesian calibrations allow both model structure and parameters to be assessed simultaneously in a probabilistic framework, providing a model with which forecasts and their uncertainty can be better understood and quantified using posterior probability distributions. Two different structures for including local productivity in the model were compared with a Bayesian model comparison. A complete calibration of the more probable model structure was then completed. Example forecasts from the calibration were compatible with existing yield tables for both species. This method could be applied to other species or other model structures in the future. Finally, data assimilation was investigated as a way of reducing forecast uncertainty. Data assimilation assumes that neither observations nor models provide a perfect description of a system, but combining them may provide the best estimate. SLeDG model predictions and LiDAR measurements for sub-compartments within Queen Elizabeth Forest Park were combined with an Ensemble Kalman Filter. Uncertainty was reduced following the second data assimilation in all of the state variables. However, errors in stand delineation and estimated stand yield class may have caused observational uncertainty to be greater thus reducing the efficacy of the method for reducing overall uncertainty.

Peatlands are multiservice ecosystems: they are the largest terrestrial store of carbon in the UK, unique habitats which provide a home for internationally important species and managed for forestry, farming and game management and shooting. This makes understanding the impact of management practices on their ecology important if they are to be sustainably managed for multi-benefits. Fire has long been used to manage peatlands in the UK to improve grazing and habitat provision for livestock and game. The effect of fire on carbon cycling in blanket bogs is of increasing concern as greenhouse gas emissions from land use is now an important management as well as political issue. Gaps however, still exist in our understanding of the controls on greenhouse emissions from blanket bogs and the impact fire may have on them both directly and indirectly by modifying vegetation composition and environmental conditions. The main objective of this research was to assess the effect of fire on greenhouse gas emissions by measuring methane and ecosystem respiration after burning at blanket bog sites across Scotland for a period of up to 3 years and relating changes in fluxes with changes in vegetation composition and abiotic conditions. In addition, the response of the Sphagnum layer to burning was assessed by looking at the recovery of Sphagnum capillifolium in the field and in a novel laboratory experiment. The indirect effects of fire on methane emissions were further investigated by a laboratory experiment devised to test if high temperatures would be fatal to methanotrophic bacteria in the Sphagnum layer, reducing methanotrophy, and thus a mechanism for fire to increase methane emissions in the short term. The results showed that methane emissions and ecosystem respiration were not significantly different in burnt plots when compared to adjacent unburnt plots at each of the three sites studies. Methane emissions were only weakly correlated to the position of the water table and neither methane fluxes or ecosystem respiration correlated with measures of vegetation composition and above ground biomass. Methanotrophy in Sphagnum was found to be difficult to detect, with a high temperature treatment having no significant effect on rates of methane oxidation. S. capillifolium was found to respond to fire by growing new auxiliary stems if the capitulum was consumed or irreversible damaged physiologically by temperatures experienced at the moss surface, with surface temperatures around 400oC with a temperature residency time of 30 seconds on artificially dried samples the most damaging, but not lethal, treatment. These results suggest that low severity fires which only consume the canopy vegetation, not penetrating the peat and leaving the moss layer mostly intact, do not have significant effects on methane emissions and ecosystem respiration in the short and medium term. In addition, it suggests that S.capillifolium can, under certain circumstances, survive a fire with the characteristics of those studied here. These findings reiterate that best practice burning guidelines must continue to ensure that burning is only carried out on blanket bog when conditions are conducive to fires with the characteristics studied here, which had little effect on important components of the carbon cycle and are survivable by at least one of the most common species of Sphagnum.

Tropical savannas and woodlands are the dominant vegetation cover in Southern Africa covering 4 million km2. Their large spatial extent means they are potentially a globally important store of biomass carbon with implications for global climate, and an area of high biodiversity value. They provide natural resources such as food, fuel and timber that help sustain the livelihoods of over 100 million people. The ability of these savanna woodlands to maintain these important ecological functions is under question due to increases in land use and land cover change. This thesis addresses a set of science questions aimed at (i) improving our knowledge of the amount of carbon and biodiversity stored in these ecosystems and how they co-vary, (ii) how these variables are spatially distributed at landscape scales and the factors which underlie these patterns, and (iii) how they respond over time to human disturbance. In Chapter 2 I examine how patterns in aboveground woody carbon storage (AGC) are linked to differences in forest structure, tree species diversity and floristic composition across a recently established network of 25 permanent sample plots in south-east Tanzania. Large stems were a significant contributor to plot-level AGC stocks with the top 3% of individuals (>40cm) in terms of size containing 35% of the total measured C. This data can potentially be used to simplify future measurements of biomass in these systems. Tree species diversity was positively related to AGC indicating the potential to align forest conservation efforts. The linear relationship suggests a functional relationship between the variables and is consistent with ecological theory on niche complementarity and selection effects, however based on the available data the mechanisms underlying this relationship can only be theorised. Changes in tree species composition were also noted across plots with differences in vegetation structure between plots explaining 16% of the variation in composition, with environmental differences related to climate and soils explaining only 3%. In Chapter 3, the focus shifts to understanding larger-scale spatial patterns in AGC. Field plots are spatially limited in this regard, therefore radar remote sensing data was used to generate a map of AGC in order to improve our knowledge on what principally controls its spatial variability at landscape scales. Results showed that factors related topography, climate and soils explained very little of the variation in C stocks across the landscape (r2 = 15 – 20%). Differences in slope angle and topographic position were important in discriminating between low biomass savannas and moderate biomass woodlands, while differences in annual precipitation were more important in separating woodlands and denser forests. A large proportion of the variation in C stocks (~80%) was unexplained highlighting the role of unmeasured variables. It is suggested that fire may play a key role in shaping patterns in tree species composition and C stocks across these landscapes. This data has important implications for a local REDD+ project which is aiming to generate carbon credits through improved fire management. In the second part of the thesis the attention shifts to understanding the long-term ecological impacts of shifting cultivation and the sensitivity and resilience of these woodlands to anthropogenic change. In Chapter 4 I examined how carbon stored in trees and soils recover across a 40-year chronosequence of abandoned agricultural land, and how this patchy disturbance impacts spatial pattern in tree species composition and diversity. I show that re-growing woodlands can act as carbon sinks through the accumulation of woody biomass (0.83 tC ha-1 yr-1), with soil texture having no clear impact on accumulation rates. Re-growing woodlands were also found to contain considerable biodiversity value by promoting novel species assemblages. Bulk soil carbon stocks appeared to be largely unaffected by the full cycle of shifting cultivation. However in Chapter 5 I show evidence of a previously unquantified legacy effect of land clearance on soil CO2 production with more recently abandoned fields (c. 6 years) exhibiting significantly higher efflux rates than the older abandonments (15 -25 years) and mature woodlands. Total soil nitrogen was the most important predictor of soil respiration across the plots (r2 = 0.3) followed by fine root density (r2 = 0.12). Soils in the younger sites were found to be more nitrogen rich which was used to explain the greater CO2 fluxes in these areas, however, it is still unclear why this pattern exists. The thesis concludes by discussing the wider implications of the results, as well as outlining further work needed to solidify some of the conclusions drawn in this thesis.

Nitrogen (N) fertilization due to atmospheric deposition (NDEP ) may explain some of the net carbon (C) sink (0.6-0.7 Pg y-1) in temperate forests, but estimates of the additional C uptake due to atmospheric N additions (∆C/∆N) can vary by over an order of magnitude (5 to 200 ∆C/∆N). High estimates from several recent studies [e.g. Magnani (2007), Nature 447 848-850], deriving ∆C/∆N from regional correlations between NDEP and measures of C uptake (such as eddy covariance -derived net ecosystem production, or forest inventory data) contradict estimates from other studies, particularly those involving 15N tracer applications added as fertilizer to the forest floor. A strong ∆C/∆N effect requires nitrogen to be efficiently acquired by trees and allocated to high C:N, long-lived woody tissues, but these isotope experiments typically report relatively little (~ 20 %) of 15N added is found above-ground, with less than 5 % of the total 15N applied found in wood. Consequently the high correlation-derived ∆C/∆N estimates are often attributed to co-variation with other factors across the range of sites investigated. However, 15N-fertilization treatments often impose considerably higher total N loads than ambient NDEP , while almost all exclusively only apply mineral 15N treatments to the soil, often in a limited number of treatment events over relatively short periods of time. Excessive N deposition loads can induce negative physiological effects and limit the resulting ∆C/∆N observed, and applying treatments to the soil ignores canopy nitrogen uptake, which has been demonstrated in numerous studies. As canopies can directly take up nitrogen, the chronic, (relatively) low levels of ambient NDEP inputs from pollution may be acquired without some of the effects of heavy N loads, with trees obtaining this N before it reaches the soil, allowing canopies to substitute for, or supplement, edaphic N nutrition. The strength of this effect depends on how much N uptake can occur across the canopy under field conditions, and if this extra N supplies growth in woody tissues such as the stem, as well as the canopy. Similarly, such mineral fertilizer isotope trace experiments are also unable to trace N in the decomposing litter and humus layers of the soil, which even under heavy NDEP loading contribute most of the N utilised for forest growth. Recent literature suggests that some organic (early decomposition) forms of N may be taken up by roots. If this litter N is not retained or distributed in the same way as mineral fertilizers, its contribution to plant nutrition and ∆C/∆N may need to be reassessed under nitrogen deposition. We tested some of these assumptions in the nursery and the field. In order to facilitate litter 15N tracing, we conducted an experiment injecting large trees with 15N-NH4NO3 to create 15N-labelled litter, tracing the applied isotope into a full harvest of the canopy. Such labelled litter substitute was used to replace the litter layer in a Sitka Spruce plantation (Picea sitchensis L. (Bong.)), where the fate of this 15N from litter decomposition in the soil system was compared against the fate of 15N in deposition. Similarly, in potted Sitka Spruce saplings, we used combination treatments of 15N-labelled litter, soil-targeted 15N-deposition, and canopy targeted 15N-deposition, investigating 15N return in different age classes of above and below ground biomass. We found that i) 15N recovery in canopies (needles and branches) in our injected trees was almost all of the injected 15N five months after injection, ii) canopy application of NDEP led to 60 % 15N return in above-ground parts of saplings compared to 21 % in soil applications and iii) a litter-derived 15N source was retained 55 % more in topsoil, and 36 % more in roots, than a similar deposition 15N source applied as mineral fertilizer. We discuss the implications of such findings in the context of 15N return in different plant organs and ecosystem pools, seasonal variation in N content, and overall inferences of a forest ∆C/∆N effect. Our results suggest that the total ∆C/∆N effect driven by a high N sequestration from canopy uptake in wood is ~ 114:1, more than double that of 15N tracer experiments but not as high as upper estimates from correlative studies, and that litter-derived organic N is better retained in trees and soils in excess of similar amounts of mineral 15N from deposition. Existing forest 15N-fertilization experiments could under-estimate the overall ∆C/∆N effect of atmospheric N deposition.

Leaf traits have long been used to classify and characterise species in natural ecosystems. In addition, leaf traits provide important information about plants’ strategies for the use of resources and can be used to improve our understanding of ecosystem level processes such as nutrient cycling and carbon allocation. To explore the linkages between leaf traits and productivity, we worked in two resource-limited ecosystems (a grassland and a forest), and used leaf traits to understand how species respond to changes in available resources and their relationship to ecosystem processes. We worked in a species rich limestone-grassland located in central England, which has been subjected to long-term climatic manipulation (winter warming, summer drought and extra summer rainfall). We characterised species composition in terms of their identity, abundance and leaf structural properties (nitrogen content and leaf mass per area (LMA)) in the main treatments and the control. We found that change in species abundance was the most important factor to understand the differences in productivity (above ground biomass and total foliar nitrogen). We then measured CO2 exchange at ecosystem level, using a chamber technique, and assessed the treatments’ effect on the gross primary productivity (GPP) and ecosystem respiration (Reco). GPP and Reco were controlled by soil moisture and above ground biomass but also influenced by the conditions experienced during the growing season prior to the measuring period. Our second location was a post-disturbance chronosequence in a seasonally dry tropical forest in Costa Rica and we used leaf level gas exchange measurements to explore the role of nitrogen (N) and phosphorus (P) on the temporal-spatial variation of photosynthesis of dominant species. We found that photosynthetic efficiency was strongly linked to leaf N and P content, but that there was an important seasonal pattern on this relationship likely associated to P remobilization. Additionally we found seasonal changes in resources (water, nutrients) had a larger impact on the photosynthetic parameters than changes along the chronosequence. The two ecosystems studied for this thesis are contrasting in their physiognomy, species composition and climate, but are also characterised by species whose structural traits (high LMA and high C:N ratio) are likely to have a significant impact on the nutrient cycling processes. We learned that leaf traits provide important information about species strategies and their usage of resources and they can also aid to address questions at ecosystem level in time and space, either through simple aggregation or as emergent properties. Additionally, the traits explored are important input information to up-scale processes from leaf to the ecosystem level, a step needed to address the effect changes in resources will have on the seasonally dry tropical forest and grasslands, which represent a significant fraction of the total global carbon storage.

In this study, theoretical and experimental approaches were combined, using Arabidopsis as the studied species. The multi-scale model incorporates the following, existing sub-models: a phenology model that can predict the flowering time of plants grown in the field, a gene circuit of the circadian clock network that regulates flowering through the photoperiod pathway, a process-based model describing carbon assimilation and resource partitioning, and a functional-structural module that determines shoot structure for light interception and root growth. First, the phenology model was examined on its ability to predict the flowering time of field plantings at different sites and seasons in light of the specific meteorological conditions that pertained. This analysis suggested that the synchrony of temperature and light cycles is important in promoting floral initiation. New features were incorporated into the phenology model that improved its predictive accuracy across seasons. Using both lab and field data, this study has revealed an important seasonal effect of night temperatures on flowering time. Further model adjustments to describe phytochrome (phy) mutants supported the findings and implicated phyB in the temporal gating of temperature-induced flowering. The improved phenology model was next linked to the clock gene circuit model. Simulation of clock mutants with different free-running periods highlighted the complex mechanism associated with daylength responses for the induction of flowering. Finally, the carbon assimilation and functional-structural growth modules were integrated to form the multi-component, whole-plant model. The integrated model was successfully validated with experimental data from a few genotypes grown in the laboratory. In conclusion, the model has the ability to predict the flowering time, leaf biomass and ecosystem exchange of plants grown under conditions of varying light intensity, temperature, CO2 level and photoperiod, though extensions of some model components to incorporate more biological details would be relevant. Nevertheless, this meso-scale model creates obvious application routes from molecular and cellular biology to crop improvement and biosphere management. It could provide a framework for whole-organism modelling to help address global issues such as food security and the energy crisis.

Amazonia contains the largest expanse of tropical forest in the world and is globally significant as a store of carbon, a regulator of climate and an area of high species diversity. The ability of the Amazonian forests to maintain these important ecological functions is however, increasingly under question in light of recent predictions of climate change. There is currently significant uncertainty in model predictions of how Amazonian forests will respond to predicted future climate change. This thesis reports the finding of two field studies, targeted at understanding the responses of two tropical forest carbon fluxes which are poorly simulated in vegetation models, and two modelling studies, which aim to better quantify uncertainty on model predictions of the effects of current and future climate change on the ecological function of Amazonian forests. The responses of forests to varying magnitudes of seasonal changes in climate which occur across Amazonia can give an important insight into the sensitivity of these forests to climate perturbations and changes. Testing the sensitivity of an Amazonian forest in Tambopata, Peru, to seasonal variations in precipitation and temperature, I find that the stem diameter growth of tropical trees is more sensitive to water availability than temperature changes. The vulnerability of trees to reduced soil water varied between tree classes with different functional traits, including wood density, tree height, tree diameter and tree growth rate. Similarly, I find that the respiration flux from tropical dead wood, at a second site in French Guiana, is highly sensitive to variations in water content. I show that these variations in respiration fluxes can be modelled successfully using seasonal variations in soil water content. To date there are few studies which have comprehensively tested vegetation models using ecological data from Amazon forests. Using data assimilation and nine sources of ecological data I estimate the certainty with which we can parameterise a carbon cycle model to represent the effects of a strong dry season on tropical forests. Using this technique I find, that the carbon balance of Amazonian forests can be very sensitive to reductions in water availability, and that these seasonal changes need to be accurately simulated across models to correctly predict annual carbon budgets. The variability in model responses caused by differences in the way processes are structured and parameterised in vegetation models requires better quantification. Using a model inter-comparison I demonstrate that the relative sensitivity of modelled climate-vegetation feedbacks to changes in ambient air temperature and precipitation is highly variable. I find that although the models showed similar directional responses at both the leaf and canopy scale some models showed a greater sensitivity to temperature and others to drought. I therefore demonstrate the need for greater constraint on modelled responses of Amazonian forests to changes in temperature and precipitation. The impact of climate change on Amazonian forests is an important global issue, yet our knowledge is reliant on our ability to understand the uncertainties on our predictions. Using field data to evaluate and to develop model predictions is a valuable way to reduce the uncertainty associated with modelling future change. This thesis presents an investigation of how tropical forests respond to changes in climate and with what certainty we can model these changes in order to predict the response of Amazon forests to predicted future climate change.

Africa’s carbon (C) cycle is one of the least well understood components of the global C cycle. Miombo woodlands are the most common woodland type in southern Africa, but despite their vast extent and importance in the biogeochemical cycles of Africa, their C dynamics are not well understood. This thesis addresses a set of science questions related to miombo woodland C dynamics that cover a range of scales, from the leaf to the landscape. The questions are related to seasonal controls on C uptake at the leaf level, to spatial distributions and scales of variation of C stocks in the landscape, and to the drivers and spatial patterns of deforestation and degradation at the regional scale. In miombo woodlands, the seasonality of productivity remains poorly understood, and it is unclear whether stomatal limitations or variations in leaf traits cause seasonal changes in productivity. I use data of leaf gas exchange and leaf traits collected in dry and wet seasons to assess the response of photosynthesis to seasonality. I found a large degree of inter-specific responses, where photosynthetic capacity was maintained between seasons in some tree species but not in others. This was linked to inter-specific stomatal regulation on leaf gas exchange, access to soil water and varied leaf traits, indicating differing timing of leaf development during the dry season. Differing timing of leaf flushing can create niche separation, facilitating the co-existence of miombo woodland tree species. I use data collected along a 5 km transect through miombo woodland to characterise the spatial distributions and scales of variation of C stocks in woody biomass and soils, and assess the links between them. I found that on the scale of a few meters, soil C stocks varied in relation to soil texture. At the kilometre scale, surface soil and woody C stocks were coupled, and varied in relation to topography. By understanding the scales of variation I was able to make recommendations for optimal sampling of C stocks in a miombo woodland landscape for improved C stock assessments. I developed and tested a simple spatial model of deforestation and degradation, using a rule-based approach, to produce risk maps of areas more likely to be affected by deforestation and degradation for a study site in central Mozambique. I found that my model was able to accurately predict the locality of high risk areas, and that roads were the major axis for forest biomass loss. Risk maps created from this method are useful for exploring the drivers of deforestation and degradation in a region dominated by miombo woodland, and for targeting policy and management efforts. Overall, this thesis has contributed significantly to our understanding of natural and human driven miombo woodland C dynamics over a range of scales, from the leaf to the landscape. In the final chapter, I discuss the implications of each chapter for our understanding of miombo woodland C dynamics, and suggest areas for further research.

The soil organic carbon (C) pool is estimated to contain at least three times as much organic C as is stored in vegetation. However, the processes controlling below-ground C dynamics are poorly understood, representing a key uncertainty in ecosystem models. Soil respiration rate (Rs) is a large component of the forest carbon cycle, however the factors that control it are still poorly understood, and those affecting autotrophic (Ra) and heterotrophic (Rh) respiration rates differ and vary in space and time. A variety of direct (i.e. soil and ingrowth cores) and indirect (i.e. rhizotron and minirhizotron) methods exist for obtaining estimates of fine root (< 2 mm diameter) production, with the consequence that there is a high variability in root biomass estimates between root studies. In this thesis I aim to contribute towards a better understanding of processes governing below-ground C dynamics. In particular I focus on: 1) the spatial and seasonal variability of Rs and drivers; 2) the uncertainty on fine root C pool measurement methods; 3) comparing novel datasets of Rs, fine root biomass and girth increment, with outputs from the SPA v2 model. To determine the dominant controls and spatial heterogeneity of Rs, I measured Rs and key biotic and abiotic drivers seasonally, in a Quercus robur forest in southern England. Measurements were made quarterly in three plots, each with measurement points arranged according to a spatial sampling design, enabling any spatial autocorrelation to be detected. Rs drivers were categorised into plant (i.e. leaf area index, weighted tree proximity (i.e. mean dbh within 4 m of a point), and fine root biomass), physical (i.e. soil moisture, soil temperature and soil bulk density) and substrate (i.e. litter depth and organic layer depth) factors. I explore: 1) what the dominant controls of Rs are and whether they change during the growing season; 2) whether micro-topography and stand structure are correlated with drivers, and influence the spatial variability of Rs, thereby simplifying up-scaling processes; 3) if physical drivers of Rs are spatially more homogeneous than plant drivers and the availability of substrate. I found no clear seasonal difference in drivers, with Rs consistently responding to litter depth, bulk density and soil moisture. The only significant response of Rs to micro-topography and tree factor was in August and September respectively and physical factors were found to be the most spatially homogeneous. Rs measurements were non-normally distributed, with ‘hotspots’ of particularly high fluxes found that remained stable throughout the measurement campaign. These findings suggest that the seasonal and spatial variability and distribution of Rs and its main drivers should be considered at the sampling design stage, to avoid bias for up-scaling non-linear processes. To address the uncertainties associated with determining fine root biomass change, we compared the measurement error for five methodologies (four indirect and one direct) in a Pinus contorta and Quercus robur forest during 2010. Rhizotron and ingrowth measurements were taken during 2010 and fine root standing crop was measured in 2009. Root length against the rhizotron screens was measured using novel software (ORIDIS), developed as part of a collaboration here in Edinburgh. The software was developed to increase precision and reduce the cost and processing time of rhizotron measurements. Differences in final cumulative root ingrowth for each conversion method ranged between 20.7g-2 - 245.0 g m-2 in the oak forest and 89.7 g m-2 - 273.0 g m-2 in the pine forest. The study found that indirect measurements of root length had less operator error than indirect measurements of root diameter. Direct methods of determining root growth using ingrowth cores also showed a seasonal trend; however artefacts may have been introduced into the method, from the affect of severing roots and changing soil conditions. To test the representation of below-ground processes in an ecosystem model, I validate modelled dynamics using default SPA v2 parameters, against independent CO2 flux and C pool datasets. The flux data were of eddy covariance and automatic chamber measurements, partitioned into root (Rroot), mycorhizal (Rmyc) and microbial heterotrophic (Rh) components. The biometric measurements were of foliage, fine root biomass and woody biomass increment. The key findings of this study were that: 1) SPA outputs compare well to ecosystem scale measurements of NEE and GPP. However, model-data mismatch occurs for fine root and wood C allocation; 2) the timing of fine root C allocation is 53 days too late and the turnover rate of fine roots 17 times too high; 3) the timing of modelled below-ground Rh and Ra could be improved by separating above and below-ground Ra and including individual root, mychorrizae and microbial C pools. The thesis concludes by discussing the implications of each chapter for our understanding and capability to model below-ground C dynamics. I find that the key challenge for measuring individual below-ground C pools and fluxes is ensuring that the measurements are spatially representative and avoid bias. The key challenge for modelling below-ground C dynamics is ensuring processes sufficiently reflect reality, when the sparse data that exist for corroboration, capture multiple processes. I explore the possibilities of further research that could be conducted, as a result of this work.

The interactions between vegetation and climate are complex and critical to our ability to predict and mitigate climate change. Savanna ecosystems, unique in their structure and composition, are particularly dynamic and their carbon cycling has been identified as highly significant to the global carbon budget. Understanding the responses of these dynamic ecosystems to environmental conditions is therefore central to both ecosystem management and scientific knowledge. Longleaf pine ecosystems are highly biodiverse and unique savanna ecosystems located in the south-eastern USA – an important current carbon sink and key area identified for future carbon sequestration. These ecosystems depend on fire to maintain their structure and function, and the longleaf pine tree itself (Pinus palustris Mill.) has been noted for its resilience to drought, fire, pests and storms and is thus becoming increasingly attractive as both a commercial forestry species and a provider of other ecosystem services. Previous process-based models tested in the south-eastern USA have been shown to fail in conditions of drought or rapid disturbance. Consequently, in order to inform management and understand better the physiology of these ecosystems, there is a need for a process-based model capable of upscaling leaf-level processes to the stand scale to predict GPP of longleaf pine savannas. P.palustris exists across a wide range of soil moisture conditions, from dry sandy well-drained soils (xeric) to claypan areas with higher moisture content (mesic). Previous work has demonstrated that this species adjusts many aspects of its physiology in response to these differing soil conditions, even under identical climate. The research in this thesis supports these previous findings, but additionally explores, with the assistance of the Soil Plant Atmosphere model (SPA), the productivity response of P. palustris across the soil moisture gradient. Contrary to expectations, measurements, field observations and modelling suggest that P. palustris trees growing in already water-limited conditions cope better with exceptional drought than their mesic counterparts. At the leaf-level, xeric P. palustris trees were found to have higher measured net photosynthesis, but the lower stand density and leaf area at this site meant that in non-drought conditions mesic P. palustris annual gross primary productivity (GPP) was 23% greater than xeric annual GPP. Initial upscaling of leaf-level processes to the canopy scale using the SPA model found that, during the growing season when other components of longleaf pine ecosystems are active, the longleaf pine may only be responsible for around 65% of the total productivity. Other important components of longleaf pine savannas are oaks and grasses which, with pine, constitute 95% of longleaf pine ecosystem biomass. Each of these groups, however, responds differently to fire and water availability. Despite this, the other components of longleaf pine savannas have received limited research attention and have never been modelled using a process-based model such as SPA. As integral components of longleaf pine carbon budgets, it is essential that the physiology and productivity of oaks and grasses in this system are better understood. The research in this thesis studied the productivity response of these groups during drought across a soil moisture gradient, and found that oak and pines at each site appear to fill separate ecohydrological niches depending on whether or not they are growing in a xeric or mesic habitat. As expected, the highest drought tolerance was found in the C4 grass, wiregrass (Aristida stricta), at both xeric and mesic sites. In order to further explore the contributions of the different functional groups in longleaf pine savannas, the SPA model was adapted to run with concurrent functional groups and to represent the different photosynthetic pathways of the understorey grasses (C4) and the canopy trees (C3). The aim of this part of the thesis was to represent better a savanna ecosystem in a process-based model and explore and quantify the contributions of each functional group diurnally, seasonally, annually and interannually. Modelling results suggest that accurately representing the phenology not only of trees but of grasses, is critical to capturing ecosystem GPP and its variability. This phenology may not only be seasonally controlled, but also dictated by fire. Overall, this research highlights the importance of continued research into savanna and savanna-like ecosystems. Additionally, it provides an insight into the responses of multiple ecosystem components to an extreme drought, and how these responses differ at leaf, stand and landscape scales. The thesis also employs a little-used method of combining eddy-covariance data with a process-based model to separate out different ecosystem components, a method becoming more common but not yet widely tested.

In the year 2000, croplands covered about 12% of the Earth’s ice-free land surface. Through cropland management, humankind momentarily appropriates about 25% of terrestrial ecosystem productivity. Not only are croplands a key element of human food supply, but also bear potential in increased carbon (C) uptake when best-practice land management approaches are adopted. A detailed assessment of the impact of land use on terrestrial ecosystems can be achieved by modelling, but the simulation of crop C cycling itself is a relatively new discipline. Observational data on crop net ecosystem exchange (NEE) are available only recently, and constitute an important tool for model development, diagnosis, and validation. Before crop functional types (CFT) had been introduced, however, large-scale biogeochemical models (BGCM) lacked crop-specific patterns of phenology, C allocation, and land management. As a consequence, the influence of cropland C cycling on biosphere-atmosphere C exchange seasonality and magnitude is currently poorly known. To date, no regional assessment of crop C cycling and yield formation exists that specifically accounts for spatially and temporally varying patterns of sowing dates within models. In this thesis, I present such an assessment for the first time. In the first step (chapter 2), I built a crop C mass balance model (SPAc) that models crop development and C allocation as a response to ambient meteorological conditions. I compared model outputs against C flux and stock observations of six different sites in Europe, and found a high degree of agreement between simulated and measured fluxes (R2 = 0.83). However, the model tended to overestimate leaf area index (LAI), and underestimate final yield. In a model comparison study (chapter 3), I found in cooperation with further researchers that SPAc best reproduces observed fluxes of C and water (owed to the model’s high temporal and process resolution), but is deficient due to a lack in simulating full crop rotations. I then conducted a detailed diagnosis of SPAc through the assimilation of C fluxes and biometry with the Ensemble Kalman Filter (EnKF, chapter 4), and identified potential model weaknesses in C allocation fractions and plant hydraulics. Further, an overestimation of plant respiration and seasonal leaf thickness variability were evident. Temporal parameter variability as a response to C flux data assimilation (DA) is indicative of ecosystem processes that are resolved in NEE data but are not captured by a model’s structure. Through DA, I gained important insights into model shortcomings in a quantitative way, and highlighted further needs for model improvement and future field studies. Finally, I developed a framework allowing for spatio-temporally resolved simulation of cropland C fluxes under observational constraints on land management and canopy greenness (chapter 5). MODIS (Moderate Resolution Imaging Spectroradiometer) data were assimilated both variationally (for sowing date estimation) and sequentially (for improved model state estimation, using the EnKF) into SPAc. In doing so, I was able to accurately quantify the multiannual (2000-2006) regional C flux and biometry seasonality of maize-soybean crop rotations surrounding the Bondville Ameriflux eddy covariance (EC) site, averaged over 104 pixel locations within the wider area. Results show that MODIS-derived sowing dates and the assimilation of LAI data allow for highly accurate simulations of growing season C cycling at locations for which groundtruth sowing dates are not available. Through quantification of the spatial variability in biometry, NEE, and net biome productivity (NBP), I found that regional patterns of land management are important drivers of agricultural C cycling and major sources of uncertainty if not appropriately accounted for. Observing C cycling at one single field with its individual sowing pattern is not sufficient to constrain large-scale agroecosystem behaviour. Here, I developed a framework that enables modellers to accurately simulate current (i.e. last 10 years) C cycling of major agricultural regions and their contribution to atmospheric CO2 variability. Follow-up studies can provide crucial insights into testing and validating large-scale applications of biogeochemical models.

Rapid climate change in Arctic regions is of concern due to important feedbacks between the Arctic land surface and the global climate system. A large amount of organic carbon (C) is currently stored in Arctic soils; if decomposition is stimulated under warmer conditions additional release of CO2 could result in an accelerating feedback on global climate. The strength and direction of Arctic C cycle - climate feedbacks will depend on the growth response of vegetation; if plant growth increases some or all of the extra CO2 emissions may be offset. Currently the Arctic is thought to be a small net sink for CO2, the expected balance of terrestrial C sinks and sources in the future is unknown. In this thesis I explore some of the critical unknowns in current understanding of C cycle dynamics in Arctic vegetation. Quantifying gross primary productivity (GPP) over regional scales is complicated by large spatial heterogeneity in plant functional type (PFT) in Arctic vegetation. I use data from five Arctic sites to test the generality of a relationship between leaf area index (LAI) and canopy total foliar nitrogen (TFN). LAI and TFN are key drivers of GPP and are tightly constrained across PFTs in Low Arctic Alaska and Sweden, therefore greatly simplifying the task of up-scaling. I use data from Greenland, Barrow and Svalbard to asses the generality of the LAI-TFN relationship in predicting GPP at higher Arctic latitudes. Arctic ecosystems are unique among biomes in the large relative contribution of bryophytes (mosses, liverworts and hornworts) to plant biomass. The contribution of bryophytes to ecosystem function has been relatively understudied and they are poorly represented in terrestrial C models. I use ground based measurements in Northern Sweden to fill an existing data gap by quantifying CO2 fluxes from bryophytes patches in early spring and summer, and develop a simple model of bryophyte GPP. Using the model I compare bryophyte GPP to that of vascular plants before, during and after the summer growing season, finding that productive bryophyte patches can contribute up to 90 % of modelled annual GPP for typical vascular plant communities at the same site, and that the relative magnitude of bryophyte GPP is greatest in spring whilst the vascular plant canopy is still developing. Understanding how GPP relates to plant growth is important in relating remotely sensed increases in Arctic ‘greenness’ to changes in plant C stocks. I use a 13C pulselabelling techniques to follow the fate of recently fixed C in mixed vascular and bryophyte vegetation, with a focus on quantifying the contribution of bryophytes to ecosystem carbon use efficiency (CUE). I show that bryophytes contribute significantly to GPP in mixed vegetation, and act to increase ecosystem CUE. I highlight the importance of including bryophytes, which do not have roots, in aboveground: belowground partitioning schemes in C models. To further explore C turnover in bryophytes, I use the results of a second 13C labelling experiment to develop a model of C turnover in two contrasting Arctic mosses (Polytrichum piliferum and Sphagnum fuscum). I find significant differences in C turnover between Polytrichum piliferum which respires or translocates about 80 % of GPP, while Sphagnum fuscum respires 60 %. This analysis is the first to explicitly model differences in C partitioning between Arctic bryophyte species. Finally, I discuss the implications of each chapter for our understanding of Arctic C dynamics, and suggest areas for further research.

The Amazon rain forest plays an important role in regional and global biogeochemical cycling, but the region may undergo an increase in the frequency and severity of drought conditions driven by global climate change, regional deforestation and fire. The effects of this drought on carbon cycling in the Amazon, particularly below-ground, are potentially large but remain poorly understood. This thesis examines the impacts of seasonal and longer-term drought upon ecosystem carbon allocation and cycling at an Amazon rain forest site with a particular focus upon below-ground processes. Measurements are made at three one-hectare forest plots with contrasting soil type and vegetation structure, to observe responses across a range of Amazon primary forest types. A fourth plot is subjected to partial rainfall exclusion to permit measurement of forest responses to a wider range of soil moisture levels than currently exists naturally. An analysis of the number of samples required to accurately quantify important ecosystem carbon stocks and fluxes is used to guide the sampling strategy at the field site. Quantifying root dynamics, in particular, presents methodological challenges. Thus, I critically review existing methods, and develop techniques to accurately measure root standing biomass and production. Subsequently, these techniques are used to record root responses, in terms of standing biomass, production, morphology, turnover and nutrient content, to variation in soil moisture across the four rain forest plots. There is substantial environmental variation in root characteristics. However, several responses remain consistent across plots: root production of biomass, length, and surface area, is lower where soil is dry, while root length and surface area per unit mass show the opposite pattern. The other major component of the below-ground carbon cycle is soil carbon dioxide efflux. I partition this efflux, on each plot, into contributions from organic ground surface litter, roots and soil organic matter, and investigate abiotic and biotic causes for observed differences within and between plots. On average, the percentage contribution of soil organic matter respiration to total soil carbon dioxide efflux declines during the dry season, while root respiration contribution displays the opposite trend. However, spatial patterns in soil respiration are not directly attributable to variation in either soil moisture or temperature. Instead, ground surface organic litter mass and root mass account for 44 % of observed spatial heterogeneity in soil carbon dioxide efflux. Finally, information on below-ground carbon cycling is combined with aboveround data, of canopy dynamics and stem wood production and mortality, to analyze the potential effects of drought upon carbon cycling in an Amazon forest ecosystem. Comparison of the rainfall exclusion plot with a similar, but unmodified, control plot reveals potentially important differences in tree carbon allocation, mortality, reproduction, soil respiration and root dynamics. The apparent net consequence of these changes is that, under drier conditions, the amount of CO2 moving out of the forest and into the atmosphere is diminished. This synthesis of above-ground and below-ground data advances understanding of carbon cycling in rain forests, and provides information which should allow more accurate modelling of the response of the Amazon region to future drought. Additional measurements at other sites, and of other ecosystem carbon fluxes, should further refine modelling predictions.