1. The relevance of studies of development in crop-weather investigations is reviewed and the aims of the present work are outlined. 2. The procedures used in studying development in field crops of winter wheat are described. The developmental progress of the plants was ascertained by frequent dissections. 3. Primordium initiation at the stem apex is strongly dependent upon apex temperature, which could be accurately estimated from standard meteorological screen temperatures. Like numerous other complex biological processes, initiation has a markedly linear response to temperature: the number of primordia initiated is therefore in direct proportion to accumulated temperature (thermal time). To calculate this requires estimation of the base temperature (Tb). 4. The linear dependence upon temperature of the initiation rates of leaves, spikelets and florets (R1, Rs and Rf ) was evident. Spikelets were initiated faster than leaves ; rate changed at a distinct inflexion point, usually at about the end of leaf initiation but sometimes later. Tb = 0 grad. C for leaves but was higher for spikelets and florets. The shift in Tb apparent17 occurred because Rs and Rf were strongly influenced by the day length at inflexion point. When temperature was corrected for day length influence, Tb = 0°0 for each developmental phase. Inflexion point timing apparently depended upon interaction between vernalisation before crop emergence and photo thermal time afterwards. 5. Leaf appearance rate in thermal time was linear but apparently influenced by the direction and magnitude of day length change at emergence, with a possible secondary effect of current day length. Leaf extension was strongly related to temperature. The gradient of lamina size up the stem appeared to be ontogenetically determined. 6. Compared with early-sown or fully-fertilised crops, floret survival and grain yield was lower in those sown late or inadequately fertilised, probably on account of their smaller amount of growth per unit of developmental time.

The research described in this thesis expresses the importance of quantitative and qualitative data types and how these can be incorporated and combined to produce an agricultural management decision support system (DSS). Researchers cannot solely depend on numerical data and relationships when designing, modelling and producing decision management tools. The relevance of the social sciences and peoples interpretations of these tools is equally important. The DSS described here focuses on the management of rainwater harvesting (RWH) in Tanzania. Numerical data related to natural resources (water and nutrients) and yields of rice and maize have been collected for the production of the DSS. With regard to the social science factors, the DSS tackles the concept of common pool resources (CPR) of water and nutrients. The importance of CPR is well understood, however their inclusion in the production of models is a relatively new concept. Criteria related to social status is linked with the by laws that govern the allocation of natural resources in Tanzania to help derive a numerical method for including CPR within the DSS. The production of the DSS is a novel way of combining this research into a tool that aims to benefit all socio-economic community groups. During the production of the DSS, a single generic approach for the inclusion of quantitative and qualitative information has developed. Particular focus was on the development of a model base (programming and mathematical relationship building), database (storage of the data used for the relationships) and a dialog system (the user-interface and communication strategy). This method is termed the ‘dialog, data, and models (DDM)’ paradigm (Sprague and Carlson, 1982). From this research, a DSS has been produced that aims to optimise RWH management in Tanzania with the aim of alleviating poverty and enhancing sustainable agriculture for all community members. Also an overall strategy for the production of DSSs has been produced. It illustrates how both quantitative (numerical and physical data) and qualitative (socio-economic considerations) can be utilised individually and in combination for the production of DSSs and can be extrapolated for further research and to new areas.

Increasing demand for timber, fuelwood and other forest products has outstripped production in many areas of the semi-arid tropics, leading to deforestation and land degradation resulting from erosion and nutrient depletion. Agroforestry offers the potential to provide forest products, improve productivity and reduce soil erosion and environmental degradation. However, as previous reports have shown that competition between trees and crops for water in semi-arid areas adversely affects crop yields, attention has turned towards studies of the existing practice of boundary planting, where low tree planting densities and the associated benefits of land demarcation and soil stabilisation increase the viability of incorporating trees into crop land. The aim is to select and manage tree species in ways that limit their negative effect on crop yields and improve the overall value of the system. The present study was carried out at Machakos (1° 33' S, 37° 14' E, altitude 1660 m) in the Kenyan highlands, where the bi-modal annual rainfall of c. 740 mm is divided approximately equally between two rainy seasons (short rains, October-February, long rains, March July). The experiment was set up in April 1993 to examine the influence of tree/crop interactions on system productivity. Each 18 x 18 m plot, except for the sole crop plots, contained a central row of trees planted at 1 m spacing. Four overstorey agroforestry treatments were examined between March 1996 and March 1998; these included two indigenous species, Croton megalocarpus and Melia volkensii, and two exotic species from Central America, Senna spectabilis and Gliricidia sepium. Beans (Phaseolusulgaris) and maize (Zea mays) were grown during the short and long rains respectively. M. volkensii and S. Spectabilis exhibited similar leafing phenology patterns, losing almost all leaf cover during the long dry season (July-October) and flushing before the ensuing rains. During the short dry season, S. spectabilis lost few leaves, whilst M. volkensii lost some leaves before flushing prior to the onset of the long rains. M. volkensii lost a large proportion of its leaf cover during the 1997/98 short rains due to the unusually high soil moisture content. C. megalocarpus although predominantly evergreen, lost a large proportion of its leaves during dry periods, whereas leaf area increased rapidly under wetter conditions. G. sepium had one annual period of low leaf cover during the long dry season and did not regain full leaf cover until mid-way through the short rains. The three-dimensional model of canopy photosynthesis and transpiration, MAESTRA, was parameterised for C. megalocarpus and M. volkensii using existing models to describe the response of photosynthesis to light and temperature and stomatal responses to light and vapour pressure deficit. The photosynthesis model fitted the experimental data well, but stomatal conductance in C. megalocarpus, although showing responses to light and vapour pressure deficit, was not closely correlated with ambient environmental conditions. M. volkensii had higher leaf area than C. megalocarpus for most of the 18 month simulation period, comprising three rainy and three dry seasons; modelled assimilation for this period was 49 % greater in M. volkensii, while canopy water use efficiency and transpiration were respectively 35 and 11 % higher. These differences accounted for the more rapid growth rate and greater competition with adjacent crops associated with M. w1kenrii relative to C. megalocarpus. Above-ground woody biomass production was greatest in M. volkensii, followed by S. spectabilis, C. megalocarpus and G. sepium; production during the fourth and fifth years after planting ranged between 2.8 and 4.9 t ha-¹ yr¹. Crop production in the agroforestry treatments was always lower than in sole crops due to below-ground competition for water and, in seasons with higher water availability, shading by the trees. Of the agroforestry systems examined, seed production for beans was highest under M. volkensii and G. sepium, followed by C. megalocarpus and S. spectabilis. Grain production in maize was greatest under C. megalocarpus, followed by G. sepium, S. spectabils and M. volkensii. Mean annual aboveground biomass production including maize grain and stover, bean seed, woody biomass and tree leaves in the M. volkensii treatment exceeded that for the sole crop plots, even though rainfall during 1996 and 1997 was only 61 and 95 % of the long term average. Although the biomass production of leaves was not estimated for S. spectabilis and G. sepium, the results obtained suggested that biomass production was greater than that obtained under sole crop cropping. The inverse correlation between tree and crop yield suggests that the value of the tree products must exceed the associated crop losses if benefits are to be obtained from agroforestry. M. volkensii is valued in areas of Kenya where markets for its products exist and therefore shows great promise for extension in semi-arid areas; where necessary, pruning may be used to reduce competition with crops and increase the length of clear bole. C. megalocarpus is widely used as a shade tree in East Africa and seems well suited for this purpose as its impact on adjacent crops was least of all the tree species examined. S. spectabilis, although having straight unbranched stems, exhibited a level of competition with adjacent crops that would necessitate a high value for its timber products to warrant its adoption. The least suitable tree species of those examined was G. sepium, whose poor form and susceptibility to attack by fungal pathogens and insects severely undermined its potential value for use in agroforestry systems.

Environmental computer models are deterministic models devoted to predict several environmental phenomena such as air pollution or meteorological events. Numerical model output is given in terms of averages over grid cells, usually at high spatial and temporal resolution. However, these outputs are often biased with unknown calibration and not equipped with any information about the associated uncertainty. Conversely, data collected at monitoring stations is more accurate since they essentially provide the true levels. Due the leading role played by numerical models, it now important to compare model output with observations. Statistical methods developed to combine numerical model output and station data are usually referred to as data fusion. In this work, we first combine ozone monitoring data with ozone predictions from the Eta-CMAQ air quality model in order to forecast real-time current 8-hour average ozone level defined as the average of the previous four hours, current hour, and predictions for the next three hours. We propose a Bayesian downscaler model based on first differences with a flexible coefficient structure and an efficient computational strategy to fit model parameters. Model validation for the eastern United States shows consequential improvement of our fully inferential approach compared with the current real-time forecasting system. Furthermore, we consider the introduction of temperature data from a weather forecast model into the downscaler, showing improved real-time ozone predictions. Finally, we introduce a hierarchical model to obtain spatially varying uncertainty associated with numerical model output. We show how we can learn about such uncertainty through suitable stochastic data fusion modeling using some external validation data. We illustrate our Bayesian model by providing the uncertainty map associated with a temperature output over the northeastern United States.

No description available.

The topic of this work concerns nonparametric permutation-based methods aiming to find a ranking (stochastic ordering) of a given set of groups (populations), gathering together information from multiple variables under more than one experimental designs. The problem of ranking populations arises in several fields of science from the need of comparing G>2 given groups or treatments when the main goal is to find an order while taking into account several aspects. As it can be imagined, this problem is not only of theoretical interest but it also has a recognised relevance in several fields, such as industrial experiments or behavioural sciences, and this is reflected by the vast literature on the topic, although sometimes the problem is associated with different keywords such as: "stochastic ordering", "ranking", "construction of composite indices" etc., or even "ranking probabilities" outside of the strictly-speaking statistical literature. The properties of the proposed method are empirically evaluated by means of an extensive simulation study, where several aspects of interest are let to vary within a reasonable practical range. These aspects comprise: sample size, number of variables, number of groups, and distribution of noise/error. The flexibility of the approach lies mainly in the several available choices for the test-statistic and in the different types of experimental design that can be analysed. This render the method able to be tailored to the specific problem and the to nature of the data at hand. To perform the analyses an R package called SOUP (Stochastic Ordering Using Permutations) has been written and it is available on CRAN.

Changepoint analysis is a well established area of statistical research, but in the context of spatio-temporal point processes it is as yet relatively unexplored. Some substantial differences with regard to standard changepoint analysis have to be taken into account: firstly, at every time point the datum is an irregular pattern of points; secondly, in real situations issues of spatial dependence between points and temporal dependence within time segments raise. Our motivating example consists of data concerning the monitoring and recovery of radioactive particles from Sandside beach, North of Scotland; there have been two major changes in the equipment used to detect the particles, representing known potential changepoints in the number of retrieved particles. In addition, offshore particle retrieval campaigns are believed may reduce the particle intensity onshore with an unknown temporal lag; in this latter case, the problem concerns multiple unknown changepoints. We therefore propose a Bayesian approach for detecting multiple changepoints in the intensity function of a spatio-temporal point process, allowing for spatial and temporal dependence within segments. We use Log-Gaussian Cox Processes, a very flexible class of models suitable for environmental applications that can be implemented using integrated nested Laplace approximation (INLA), a computationally efficient alternative to Monte Carlo Markov Chain methods for approximating the posterior distribution of the parameters. Once the posterior curve is obtained, we propose a few methods for detecting significant change points. We present a simulation study, which consists in generating spatio-temporal point pattern series under several scenarios; the performance of the methods is assessed in terms of type I and II errors, detected changepoint locations and accuracy of the segment intensity estimates. We finally apply the above methods to the motivating dataset and find good and sensible results about the presence and quality of changes in the process.

The recent advent of Next-generation sequencing technologies has revolutionized the way of analyzing the genome. This innovation allows to get deeper information at a lower cost and in less time, and provides data that are discrete measurements. One of the most important applications with these data is the differential analysis, that is investigating if one gene exhibit a different expression level in correspondence of two (or more) biological conditions (such as disease states, treatments received and so on). As for the statistical analysis, the final aim will be statistical testing and for modeling these data the Negative Binomial distribution is considered the most adequate one especially because it allows for "over dispersion". However, the estimation of the dispersion parameter is a very delicate issue because few information are usually available for estimating it. Many strategies have been proposed, but they often result in procedures based on plug-in estimates, and in this thesis we show that this discrepancy between the estimation and the testing framework can lead to uncontrolled first-type errors. We propose a mixture model that allows each gene to share information with other genes that exhibit similar variability. Afterwards, three consistent statistical tests are developed for differential expression analysis. We show that the proposed method improves the sensitivity of detecting differentially expressed genes with respect to the common procedures, since it is the best one in reaching the nominal value for the first-type error, while keeping elevate power. The method is finally illustrated on prostate cancer RNA-seq data.

Spatial prediction of hourly rainfall via radar calibration is addressed. The change of support problem (COSP), arising when the spatial supports of different data sources do not coincide, is faced in a non-Gaussian setting; in fact, hourly rainfall in Emilia-Romagna region, in Italy, is characterized by abundance of zero values and right-skeweness of the distribution of positive amounts. Rain gauge direct measurements on sparsely distributed locations and hourly cumulated radar grids are provided by the ARPA-SIMC Emilia-Romagna. We propose a three-stage Bayesian hierarchical model for radar calibration, exploiting rain gauges as reference measure. Rain probability and amounts are modeled via linear relationships with radar in the log scale; spatial correlated Gaussian effects capture the residual information. We employ a probit link for rainfall probability and Gamma distribution for rainfall positive amounts; the two steps are joined via a two-part semicontinuous model. Three model specifications differently addressing COSP are presented; in particular, a stochastic weighting of all radar pixels, driven by a latent Gaussian process defined on the grid, is employed. Estimation is performed via MCMC procedures implemented in C, linked to R software. Communication and evaluation of probabilistic, point and interval predictions is investigated. A non-randomized PIT histogram is proposed for correctly assessing calibration and coverage of two-part semicontinuous models. Predictions obtained with the different model specifications are evaluated via graphical tools (Reliability Plot, Sharpness Histogram, PIT Histogram, Brier Score Plot and Quantile Decomposition Plot), proper scoring rules (Brier Score, Continuous Rank Probability Score) and consistent scoring functions (Root Mean Square Error and Mean Absolute Error addressing the predictive mean and median, respectively). Calibration is reached and the inclusion of neighbouring information slightly improves predictions. All specifications outperform a benchmark model with incorrelated effects, confirming the relevance of spatial correlation for modeling rainfall probability and accumulation.

The aim of this project was to develop azadirachtin as an insecticide that is applied to the soil, using a granular formulation, for root uptake and subsequent systemic plant protection. A method was developed whereby azadirachtin could be rapidly isolated to approximately 95% purity using flash chromatography. This material was used in all subsequent chemical and biochemical studies. To increase the speed of crude extract analysis, a colorimetric technique was assessed to rapidly quantify azadirachtin. However, this method was generally unsuitable for the requirements of this project because it was non-specific and not stable. Granular formulations based on sodium alginate, starch-kaolin and poly(e-caprolactone), and containing different neem seed extracts were successfully prepared. These granules exhibited differnces in the rate of azadirachtin release into water. Additives such as kaolin clay and rapeseed oil could be used to modify the speed of release. Following application to soil, the position of granules did not affect release rates. However, granule application method was shown to affect the rate at which the limonoid was accumulated within the nasturtium plants. Azadirachtin was shown to be moderately water-soluble (1.29 g/l). During mixing studies between distilled water and n-octanol, the limonoid partitioned more favourably into the non-aqueous phase at a ratio of 7:1. Based on calculated Koc values (<40), azadirachtin was classified as very highly soil mobile. Adsorption occurred principally to the organic matter of soils. Clay minerals were comparably non-sorbent. Desorption from both of these sites occurred readily. Azadirachtin was not persistent within soil where the limonoid’s DT50 was as short as 1.06 days. Initial breakdown resulted in the acetyl moiety being cleaved from the molecule. In addition, azadirachtin was shown to exhibit a pH sensitive hydrolytic degradation. The limonoid’s half-life in solution ranged from 57 days at pH 5 to 7.15 hours at pH 9. In conclusion a suitable granule for a controlled-release of azadirachtin was developed.