Modelle und Daten sind die beiden grundlegenden Elemente in den meisten Finanzmarktstudien. Viele Arbeiten konzentrieren sich auf die Verbesserung von Modellen zur besseren Annäherung an wahre Marktmechanismen, dabei konzentriert sich ein wichtiger Teil der Literatur auf die Nutzung von Informationen aus verschiedenen Quellen. In letzter Zeit haben immer mehr Forscher die Bedeutung der Modellierung aus realen Daten erkannt, dies geht einhermit der Weiterentwicklung moderner statistischer Modelle, insbesondere dem maschinellen (statistischen) Lernen, wie z. B. rekurrente neuronale Netze, die sich in den letzten Jahren bei vielen Problemen als wirksam erwiesen haben. Es hat sich gezeigt, dass der zunehmende Trend auf innovative Datenquellen wie Textnachrichten und Satellitenbilder zuzugreifen und diese zu analysieren, sich schnell zu einer wichtigen Säule der Finanzwissenschaft entwickelt hat. Auf der anderen Seite bietet die klassische Finanzliteratur eine fundierte Basis, um die aus diesen hochentwickelten Modellen und Daten gewonnenen Ergebnisse zu hinterfragen. Basierend auf der Finanzmarktanalyse mit modernen statistischen Modellen werden in dieser Dissertation in den ersten drei Kapiteln verschiedene Themen behandelt, darunter das Portfoliomanagement in Verbindung mit Informationen aus Nachrichtennetzwerken, das Risikomanagement des aufstrebenden Bitcoin-Marktes und die Vorhersage von Zeitreihen von Stromlasten mit fortgeschrittenen statistischen Modellen. / Models and data are the two fundamental elements in most of the studies on the financial market. Many papers concentrate on improving models to better approximate the true market mechanism, while an important strand of the literature focuses on exploiting more information from various sources. Recently, more and more researchers started to realize the importance of modeling from real-world data, along with the advancement of modern statistical models, especially the machine (statistical) learning models such as Recurrent Neural Network being proved to be effective on many problems in the past few years. Hence, we saw that an uprising trend of accessing and analyzing innovative data sources, such as textual news and satellite image, has been growing fast into a major pillar in financial studies. On the other hand, the classical finance literature provides us an angle to scrutinize the results generated from those sophisticated models and data. Under the spirit of financial market analysis with modern statistical models, this dissertation is written to cover various topics, including portfolio management coupled with the information from networks of news, risk management of the emerging Bitcoin market, and electricity load time series forecasting with the advanced statistical models, in the next three chapters.

This thesis presents a new template for multidisciplinary habitat mapping that combines the analyses of seafloor geomorphology, oceanographic proxies and modelling of associated biologic features. High resolution swath bathymetry of the Faial and western Pico shelves is used to present the first state-of-the-art geomorphologic assessment of submerged island shelves in the Azores. Solid seafloor structures are described in previously unreported detail together with associated volcanic, tectonic and erosion processes. The large sedimentary expanses identified in the area are also investigated and the large bedforms identified are discussed in view of new data on the local hydrodynamic conditions. Coarse-sediment zones of types hitherto unreported for volcanic island shelves are described using swath data and in situ imagery together with sub-bottom profiles and grainsize information. The hydrodynamic and geological processes producing these features are discussed. New oceanographic information extracted from satellite imagery is presented including yearly and seasonal sea surface temperature and chlorophyll-a concentration fields. These are used as proxies to understand the spatio-temporal variability of water temperature and primary productivity in the immediate island vicinity. The patterns observed are discussed, including onshore-offshore gradients and the prevalence of colder/more productive waters in the Faial-Pico passage and shelf areas in general. Furthermore, oceanographic proxies for swell exposure and tidal currents are derived from GIS analyses and shallow-water hydrographic modelling. Finally, environmental variables that potentially regulate the distribution of benthic organisms (seafloor nature, depth, slope, sea surface temperature, chlorophyll-a concentration, swell exposure and maximum tidal currents) are brought together and used to develop innovative statistical models of the distribution of six macroalgae taxa dominant in the infralittoral (articulated Corallinaceae, Codium elisabethae, Dictyota spp., Halopteris filicina, Padina pavonica and Zonaria tournefortii). Predictive distributions of these macroalgae are spatialized around Faial island using ordered logistic regression equations and raster fields of the explanatory variables found to be statistically significant. This new approach represents a potentially highly significant step forward in modelling benthic communities not only in the Azores but also in other oceanic island shelves where the management of benthic species and biotopes is critical to preserve ecosystem health.

In this paper, I analyze the effectiveness of state level Earned Income Tax Credit programs on improving of poverty levels. I conducted this analysis for the years 1991 through 2011 using a panel data model with fixed effects. The main independent variables of interest were the state and federal EITC rates, minimum wage, gross state product, population, and unemployment all by state. I determined increases to the state EITC rates provided only a slight decrease to both the overall white below-poverty population and the corresponding white childhood population under 18, while both the overall and the under-18 black population for this category realized moderate decreases in their poverty rates for the same time period. I also provide a comparison of the effectiveness of the state level EITCs and minimum wage at the state level over the same time period on these select demographic groups.

La banque de données IMIS (Integrated Management Information System) de l’agence américaine OSHA (Occupational Safety and Health Administration) contient l’ensemble des mesures de l’exposition effectuées par les inspecteurs d’OSHA chargés de vérifier la conformité aux valeurs limites d’exposition. Les résultats analytiques correspondant aux prélèvements effectués par les inspecteurs sont également disponibles dans la banque CEHD (Chemical Exposure Health Data). Ces deux banques représentent une source d’information potentielle majeure sur les conditions d’exposition aux substances chimiques en Amérique du Nord. Cependant, leur représentativité par rapport à la distribution réelle des niveaux d’exposition retrouvés dans les milieux de travail est largement inconnue. L’objectif de cette thèse est d’établir dans quelle mesure les données de contamination de l'air recueillies par l’agence fédérale américaine OSHA peuvent être utilisées pour l’estimation des expositions professionnelles en Amérique du Nord. Les analyses ont porté sur 511 047 et 588 818 mesures d’exposition contenues dans les banques IMIS et CEHD respectivement, pour la période 1979-2011. Premièrement, des modèles additifs généralisés ont été utilisés pour étudier l’association entre les variables reflétant les caractéristiques des établissements visités et des inspections et les niveaux d’exposition pour 77 agents chimiques (90% du contenu d’IMIS). Dans un second temps, une approche de régression de Poisson modifiée a été utilisée pour étudier les facteurs déterminants l’enregistrement ou non des échantillons de CEHD dans la banque IMIS en jumelant les deux banques pour 78 agents chimiques. Finalement, des modèles CART (Classification And Regression Tree) ont été développés permettant de prédire, parmi les résultats non détectés de la banque IMIS, lesquels correspondent à des mesures courte durée ou des moyennes pondérées sur 8 heures (VEMP-8h) en se basant sur les variables communes aux banques IMIS et CEHD. Dans la première analyse, les modèles statistiques ont montré que les niveaux d’exposition étaient plus susceptibles de dépasser la TLV (threshold limit value) pour les mesures effectuées sous un régime OSHA fédéral par rapport au régime OSHA d’État (rapport de cote (RC) de 1,22 à travers les agents). La probabilité de dépasser la TLV augmentait avec le nombre total des amendes reçues par un établissement, indépendamment de la nature des infractions (RC de 1,54 à travers les agents entre les catégories « élevée » et « aucune »). Elle était également plus élevée pour les visites de suivi que pour les visites planifiées (RC de 1,61). Dans la deuxième analyse, la comparaison des banques IMIS et CEHD a montré un taux d’enregistrement global de 38% des données CEHD dans IMIS. Les résultats non détectés (particulièrement ceux mesurés sur un panel d’agents – p. ex. panel de métaux) étaient moins susceptibles d’être enregistrés dans IMIS (risque relatif ~0,6). Finalement, les modèles CART ont prédit plus précisément le type de prélèvement (courte durée, VEMP-8h) pour les résultats non détectés dans IMIS que des méthodes simples d’attribution (p. ex. attribution du type le plus fréquent parmi les résultats détectés) pour les agents les plus pertinents (c.-à-d. ceux ayant une proportion substantielle de mesures ND, courte durée et VEMP-8h). Nos résultats ont montré la présence de plusieurs mécanismes de sélection dans le processus conduisant à l’enregistrement d’une mesure d’exposition dans IMIS, ce qui suggère l’existence de différences systématiques entre les niveaux rapportés dans les banques OSHA et les niveaux moyens d’exposition dans la population de travailleurs. La prise en compte des informations contextuelles aux mesures et l’emploi de méthodes prédictives peuvent aider à pallier partiellement ces biais et ainsi raffiner les portraits d’exposition établis à partir des données d’OSHA. / The Integrated Management Information System (IMIS) contains exposure measurements taken by the U.S. Occupational Safety and Health Administration (OSHA) inspectors to verify compliance with permissible exposure limits. Supplementary data containing analytical results of the field samples are available in the Chemical Exposure Health Database (CEHD). These databanks represent a major potential source of information on exposure conditions in North American workplaces. However, the degree to which they represent the actual distribution of the exposure levels found in the workplace is largely unknown.The objective of this thesis is to examine the extent to which exposure data collected by OSHA can be used for estimating occupational exposure in North America. Analyses focused on 511 047 and 588 818 exposure measurements in IMIS and CEHD respectively, for the period 1979-2011. First, generalized additive models were used to explore associations between exposure levels in IMIS and ancillary variables reflecting characteristics of establishments and inspections for 77 chemical agents (90% of IMIS content). Second, modified Poisson regression was used to identify determinants of recording or not of CEHD samples in IMIS by linking both databanks for 78 agents. Finally, Classification And Regression Tree (CART) models were applied to predict which non-detected (ND) results stored in IMIS are 8-hour time-weighted average (TWA) or short-term samples, based on common variables available in IMIS and CEHD databanks. In the first analysis, statistical modelling showed that measurements collected under federal OSHA plans were more likely to have a sample result exceed the TLV compared to measurements collected under state OSHA plans (odds ratio (OR) of 1,22 across agents). An increase in the total amount of penalty assessed to a company was associated with higher odds of having a sample result exceed the TLV (OR of 1,54 across agents for « high » vs. « none »). Follow-up inspections were more likely to have a sample result exceed the TLV compared to planned inspections (OR of 1,61 across agents). In the second analysis, linkage between CEHD and IMIS showed a 38% overall proportion of CEHD samples recorded into IMIS. Non-detects (especially ND records corresponding to analytical panels – e.g. panel of metals) were less likely to be recorded in IMIS (relative risk ~0,6). Finally, CART models predicted more accurately which IMIS ND results were TWA or short-term samples compared to simple methods of assignment (e.g. assignment of the most frequent category from detected values) for the most relevant agents (i.e. with high proportions of ND, short-term, and TWA results). Our findings showed the presence of several selection mechanisms in the process leading up to the recording of a sample in IMIS, which suggest systematic differences exist between OSHA measurements and actual occupational exposures in the general U.S. working population. These biases can be partially controlled by using ancillary information on exposure measurements together with predictive methods, thus helping to draw more accurate portraits of exposure levels from OSHA data.

O trabalho teve como foco o estudo da balneabilidade em duas praias de rios do Estado de São Paulo Balneário Reino das Águas Claras no Rio Piracuama (Pindamonhangaba, UGRHI 2) e Cachoeira das Emas no Rio Mogi-Guaçu (Pirassununga, UGRHI 9) buscando correlacionar os níveis de concentração de indicadores fecais na água com níveis de precipitação próximos e a montante da praia. O objetivo principal foi desenvolver e aplicar ferramentas de predição das condições de balneabilidade, buscando analisar e comparar os resultados obtidos nas duas praias de estudo. Essas ferramentas baseiam-se em Níveis Limiares de Precipitação (Rain Threshold Levels) e modelos estatísticos, conforme metodologias descritas principalmente em estudos publicados pela United States Environmental Protection Agency (USEPA) e pela United States Geological Survey (USGS). Foram utilizados dados das redes de monitoramento da Companhia Ambiental do Estado de São Paulo (CETESB) e do Departamento de Águas e Energia do Estado de São Paulo (DAEE), do início de 2009 ao final de 2012. Em cada praia, foram desenvolvidas ferramentas de predição para aplicação no período seco (abril a setembro) e no período chuvoso (outubro a março), com base nos níveis de precipitação acumulada em 24, 48 e 72 horas que apresentaram correlação significativa com os níveis de indicadores fecais na água. Foi considerado um período de validação (ano de 2012) independente à calibração (2009 a 2011) dos modelos. As ferramentas desenvolvidas para a praia de Cachoeira das Emas mostraram-se mais complexas do que as da praia de Balneário Reino das Águas Claras, por estar inserida em bacia hidrográfica de maiores dimensões e considerar dados de várias estações pluviométricas nos modelos de regressão. As correlações significantes (p<0,05) verificadas entre as concentrações de indicadores fecais na água e os diferentes níveis acumulados de precipitação foram positivas em ambas as praias. Pelo fato das redes de monitoramento possuir objetivos distintos, nem todos os eventos de precipitação que ocorreram durante o período de estudo puderam ser relacionados com um valor de concentração de indicador fecal na água, pois resultados microbiológicos pareados com níveis nulos ou insignificantes de precipitação não foram considerados, justamente pelo fato das ferramentas serem baseadas em chuva. As diferentes ferramentas desenvolvidas apresentaram níveis variados de especificidade e sensibilidade, ou seja, a capacidade de prever corretamente as situações de boa e má condição de balneabilidade, respectivamente. Concluiu-se que existe potencial para aplicação dessas ferramentas, porém é preciso um esforço de monitoramento conjunto dos parâmetros modelados para obter a quantidade necessária de dados necessários à boa calibração e validação, o que, consequentemente, demandaria recursos adicionais. O cenário ideal para empregá-las seria em praias bastante frequentadas num determinado período do ano, sendo a principal fonte de poluição identificada como difusa, consequente do escoamento de drenagem. A fim de melhorar seu desempenho, é possível empregar outras variáveis ambientais além de níveis de precipitação, como por exemplo, a turbidez, a vazão e o nível dos rios, desde que sejam monitorados sistematicamente junto com as condições de balneabilidade. / This study was focused on the recreational water quality at two beaches located on rivers in the State of São Paulo, Brazil Balneário Reino das Águas Claras at Piracuama River (City of Pindamonhangaba, UGRHI 2) and Cachoeira das Emas at Mogi-Guacú River (City of Pirassununga, UGRHI 9) in order to correlate concentration levels of faecal indicators in water with rainfall levels near the beach and upstream in the watershed. The main objective was the development and evaluation of predictive tools for bathing water conditions, seeking to analyze and compare the results obtained for both beaches considered in the study. These tools are based on rain threshold levels and statistical models, which methodologies are described mainly in studies published by the United States Environmental Protection Agency (USEPA) and the United States Geological Survey (USGS). Data used were collected from the monitoring networks of the State of São Paulo environmental agency (CETESB) and department of water and energy (DAEE), from 2009 to 2012. At each beach, predictive tools were developed for the dry season (April to September) and the rainy season (October to March), based on levels of accumulated rainfall in 24, 48 and 72 hours significantly correlated with concentrations of faecal indicators in water. Model development considered a validation period (year of 2012) independent from the calibration period (2009-2011). Predictive tools developed for the Cachoeira das Emas were more complex than those for the Balneário Reino das Águas Claras, because the beach is inserted into a larger watershed and its regression models considered data from various rainfall stations. The significant correlations (p value<0,05) observed between concentrations of faecal indicators in water and accumulated levels of rainfall were positive for both beaches. Because the monitoring networks have different objectives, not all rainfall events that occurred during the period of the study were related to a concentration value of faecal indicator in the water, as microbiological results paired with null or insignificant levels of precipitation were not considered, precisely because the predictive tools are based on rainfall. The different tools developed showed varying levels of specificity and sensitivity, i.e., the ability to correctly predict good and bad conditions of bathing. The conclusion is that there is a potential for application of predictive tools, however, additional effort is required on monitoring systematically the models parameters in order to obtain the necessary amount of data to achieve a good calibration and validation of the models; consequently, additional resources are required. The ideal scenario would be using the model during a certain period of the year, at beaches quite frequented, for which diffuse pollution is the main contribution, such as the drainage runoff. Aiming to improve the performance of this type of tool, one can also employ other environmental variables besides precipitation levels, such as turbidity, flow rate and river level, as long as these parameters are systematically monitored with the bathing water conditions.

O trabalho teve como foco o estudo da balneabilidade em duas praias de rios do Estado de São Paulo Balneário Reino das Águas Claras no Rio Piracuama (Pindamonhangaba, UGRHI 2) e Cachoeira das Emas no Rio Mogi-Guaçu (Pirassununga, UGRHI 9) buscando correlacionar os níveis de concentração de indicadores fecais na água com níveis de precipitação próximos e a montante da praia. O objetivo principal foi desenvolver e aplicar ferramentas de predição das condições de balneabilidade, buscando analisar e comparar os resultados obtidos nas duas praias de estudo. Essas ferramentas baseiam-se em Níveis Limiares de Precipitação (Rain Threshold Levels) e modelos estatísticos, conforme metodologias descritas principalmente em estudos publicados pela United States Environmental Protection Agency (USEPA) e pela United States Geological Survey (USGS). Foram utilizados dados das redes de monitoramento da Companhia Ambiental do Estado de São Paulo (CETESB) e do Departamento de Águas e Energia do Estado de São Paulo (DAEE), do início de 2009 ao final de 2012. Em cada praia, foram desenvolvidas ferramentas de predição para aplicação no período seco (abril a setembro) e no período chuvoso (outubro a março), com base nos níveis de precipitação acumulada em 24, 48 e 72 horas que apresentaram correlação significativa com os níveis de indicadores fecais na água. Foi considerado um período de validação (ano de 2012) independente à calibração (2009 a 2011) dos modelos. As ferramentas desenvolvidas para a praia de Cachoeira das Emas mostraram-se mais complexas do que as da praia de Balneário Reino das Águas Claras, por estar inserida em bacia hidrográfica de maiores dimensões e considerar dados de várias estações pluviométricas nos modelos de regressão. As correlações significantes (p<0,05) verificadas entre as concentrações de indicadores fecais na água e os diferentes níveis acumulados de precipitação foram positivas em ambas as praias. Pelo fato das redes de monitoramento possuir objetivos distintos, nem todos os eventos de precipitação que ocorreram durante o período de estudo puderam ser relacionados com um valor de concentração de indicador fecal na água, pois resultados microbiológicos pareados com níveis nulos ou insignificantes de precipitação não foram considerados, justamente pelo fato das ferramentas serem baseadas em chuva. As diferentes ferramentas desenvolvidas apresentaram níveis variados de especificidade e sensibilidade, ou seja, a capacidade de prever corretamente as situações de boa e má condição de balneabilidade, respectivamente. Concluiu-se que existe potencial para aplicação dessas ferramentas, porém é preciso um esforço de monitoramento conjunto dos parâmetros modelados para obter a quantidade necessária de dados necessários à boa calibração e validação, o que, consequentemente, demandaria recursos adicionais. O cenário ideal para empregá-las seria em praias bastante frequentadas num determinado período do ano, sendo a principal fonte de poluição identificada como difusa, consequente do escoamento de drenagem. A fim de melhorar seu desempenho, é possível empregar outras variáveis ambientais além de níveis de precipitação, como por exemplo, a turbidez, a vazão e o nível dos rios, desde que sejam monitorados sistematicamente junto com as condições de balneabilidade. / This study was focused on the recreational water quality at two beaches located on rivers in the State of São Paulo, Brazil Balneário Reino das Águas Claras at Piracuama River (City of Pindamonhangaba, UGRHI 2) and Cachoeira das Emas at Mogi-Guacú River (City of Pirassununga, UGRHI 9) in order to correlate concentration levels of faecal indicators in water with rainfall levels near the beach and upstream in the watershed. The main objective was the development and evaluation of predictive tools for bathing water conditions, seeking to analyze and compare the results obtained for both beaches considered in the study. These tools are based on rain threshold levels and statistical models, which methodologies are described mainly in studies published by the United States Environmental Protection Agency (USEPA) and the United States Geological Survey (USGS). Data used were collected from the monitoring networks of the State of São Paulo environmental agency (CETESB) and department of water and energy (DAEE), from 2009 to 2012. At each beach, predictive tools were developed for the dry season (April to September) and the rainy season (October to March), based on levels of accumulated rainfall in 24, 48 and 72 hours significantly correlated with concentrations of faecal indicators in water. Model development considered a validation period (year of 2012) independent from the calibration period (2009-2011). Predictive tools developed for the Cachoeira das Emas were more complex than those for the Balneário Reino das Águas Claras, because the beach is inserted into a larger watershed and its regression models considered data from various rainfall stations. The significant correlations (p value<0,05) observed between concentrations of faecal indicators in water and accumulated levels of rainfall were positive for both beaches. Because the monitoring networks have different objectives, not all rainfall events that occurred during the period of the study were related to a concentration value of faecal indicator in the water, as microbiological results paired with null or insignificant levels of precipitation were not considered, precisely because the predictive tools are based on rainfall. The different tools developed showed varying levels of specificity and sensitivity, i.e., the ability to correctly predict good and bad conditions of bathing. The conclusion is that there is a potential for application of predictive tools, however, additional effort is required on monitoring systematically the models parameters in order to obtain the necessary amount of data to achieve a good calibration and validation of the models; consequently, additional resources are required. The ideal scenario would be using the model during a certain period of the year, at beaches quite frequented, for which diffuse pollution is the main contribution, such as the drainage runoff. Aiming to improve the performance of this type of tool, one can also employ other environmental variables besides precipitation levels, such as turbidity, flow rate and river level, as long as these parameters are systematically monitored with the bathing water conditions.

Newsvendor Models with Monte Carlo Sampling by Ijeoma Winifred Ekwegh The newsvendor model is used in solving inventory problems in which demand is random. In this thesis, we will focus on a method of using Monte Carlo sampling to estimate the order quantity that will either maximizes revenue or minimizes cost given that demand is uncertain. Given data, the Monte Carlo approach will be used in sampling data over scenarios and also estimating the probability density function. A bootstrapping process yields an empirical distribution for the order quantity that will maximize the expected profit. Finally, this method will be used on a newsvendor example to show that it works in maximizing profit.

Proteins are an important component of living organisms, composed of one or more polypeptide chains, each containing hundreds or even thousands of amino acids of 20 standard types. The structure of a protein from the sequence determines crucial functions of proteins such as initiating metabolic reactions, DNA replication, cell signaling, and transporting molecules. In the past, proteins were considered to always have a well-defined stable shape (structured proteins), however, it has recently been shown that there exist intrinsically disordered proteins (IDPs), which lack a fixed or ordered 3D structure, have dynamic characteristics and therefore, exist in multiple states. Based on this, we extend the mapping of protein sequence not only to a fixed stable structure but also to an ensemble of protein conformations, which help us explain the complex interaction within a cell that was otherwise obscured. The objective of this dissertation is to develop effective ab initio methods and tools for protein un/structure prediction by developing effective statistical energy function, conformational search method, and disulfide connectivity patterns predictor. The key outcomes of this dissertation research are: i) a sequence and structure-based energy function for structured proteins that includes energetic terms extracted from hydrophobic-hydrophilic properties, accessible surface area, torsion angles, and ubiquitously computed dihedral angles uPhi and uPsi, ii) an ab initio protein structure predictor that combines optimal energy function derived from sequence and structure-based properties of proteins and an effective conformational search method which includes angular rotation and segment translation strategies, iii) an SVM with RBF kernel-based framework to predict disulfide connectivity pattern, iv) a hydrophobic-hydrophilic property based energy function for unstructured proteins, and v) an ab initio conformational ensemble generator that combines energy function and conformational search method for unstructured proteins which can help understand the biological systems involving IDPs and assist in rational drugs design to cure critical diseases such as cancer or cardiovascular diseases caused by challenging states of IDPs.

Diabetes is a major risk factor for the development of cataract [3,14,20,22]. In this thesis, we create a model that allows us to understand the incidence of one disease in the context of another; in particular, cataract in the presence of diabetes. The World Health Organization's Vision 2020 blindness-prevention initiative administers surgeries to remove cataracts, the leading cause of blindness worldwide [24]. One of the geographic areas most impacted by cataract-related blindness is Sub-Saharan Africa. In order to plan the number of surgeries to administer, the World Health Organization uses data on cataract prevalence. However, an estimation of the incidence of cataract is more useful than prevalence data for the purpose of resource planning. In 2012, Dray and Williams developed a method for estimating incidence based on prevalence data [5]. Incidence estimates can be further refined by considering associated risk factors such as diabetes. We therefore extend the Dray and Williams model to include diabetes prevalence when calculating cataract incidence estimates. We explore two possible approaches to our model construction, one a detailed extension, and the other, a simplification of that extension. We provide a discussion comparing the two approaches.

From the Proceedings of the 1971 Meetings of the Arizona Section - American Water Resources Assn. and the Hydrology Section - Arizona Academy of Science - April 22-23, 1971, Tempe, Arizona / Statistical analyses of existing hydrologic records suffer from the problem that such records are of relatively short duration, and therefore may not necessarily be random samples of the infinite population of events. On the hypothesis that tree-ring series and runoff series respond to a common climatic signal or signals that permit prediction of annual runoff from annual ring-width index, tree-ring data are used to extend available runoff records backwards in time to permit more accurate estimates of the 3 most common statistics used in hydrology: the mean, the variance and the 1st order correlation. It is assumed that both series are generated by the climatic parameters of precipitation, temperature, evapotranspiration, seasonal regime and spatial distribution. Of major concern in the reconstruction of annual runoff series from tree-ring records was the difference in persistence within each of the 2 series. A matrix of the tree-ring data was constructed, lagged up to 3 times and principal components were extracted. The covariation in this matrix was then decomposed by extracting the Eigen-vectors, and multiple regression was then used to weight the respective series and the differences in persistence were determined. This method was applied to watersheds of diverse characteristics and improved estimates of the mean and variance were obtained.