No description available.

Thesis (M.S.)--Baylor University, 2009. / Includes bibliographical references (p. 53-55).

Thesis (MScFor) -- Stellenbosch University, 2003. / ENGLISH ABSTRACT: Fire is a threat to all forest plantations. As a result, growers are forced to take active measures to reduce the incidence and extent of fires in their plantations. This thesis is an attempt to collate 846 fire records for eight Komatiland Forests (KLF) plantations in Mpumalanga North for the period 1950 to 1999. Up to now, these reports and the information therein, were not utilised by KLF for planning or for evaluating fire management practices. The only other studies in South Africa, using similar data, were conducted by LeRoux (1988) and Kromhout (1990). A brief background of the forestry industry in South Africa, and in particular Mpumalanga is presented. The main text of the report covers a presentation on fire causes, extent of damage (both in area and in Rand value) and various aspects related to time of ignition and response times. A detailed analysis was done to identify possible relationships between the variables related to compartment, climate and different fire suppression activities. A cause and frequency prediction model was developed that will assist fi re managers in identifying and determining probabilities of fires per cause. Statistical guidelines regarding the planning of fire management around fires caused by honey hunters, lightning, work-related factors, and the activities of people (public, own labour, contractors) are presented. Conclusions were drawn from the results of the analyses of the fire data, which covered a period of 47 years. Recommendations regarding guidelines for strategic fi re management for the Mpumalanga North plantations were made. The main conclusions are: • Statistics on previous fires are very useful in fire management planning as it supplies valuable information on fire causes, time of ignition , past performance related to response times, fire fighting times and damaged caused. • • • • The average area lost due to fires in the study area is 209.9 ha or 0.43% of the plantation area per annum. People-related fires (arson, smokers, picnickers, children and neighbours) caused most of the wild fires (48%), followed by lightning (22%). Some plantations performed poorly, with the occurrence of up to double the number of fires per 1 000 ha of plantation compared to other plantations in the same geographic area. There are definite patterns in the frequency of fires per cause with month of the year. These patterns are valuable for the development of strategies to manage fires caused by honey hunters, lightning fires and work-related fires. / AFRIKAANSE OPSOMMING: Brande is 'n bedreiging vir aile bosbou plantasies. Dit is dus noodsaaklik dat kwekers maatreeHs tref om die voorkoms en omvang van brande in plantasies te beperk. Hierdie tesis poog om 846 vuurverslae se inligting te ontleed ten opsigte van agt Komatiland Forests (KLF) plantasies in Mpumalanga Noord vir die tydperk 1950 tot 1999. Tot op hede is min van die inligting wat in die verslae vervat is deur KLF vir beplanning- en evalueringsdoeleindes ten opsigte van brandbestuur gebruik. Die enigste soortgelyke studies wat op brandverslagdata in Suid-Afrika gedoen is, is gedoen deur Le Roux (1988) en Kromhout (1990). 'n Kort agtergrond oor die bosbouindustrie in Suid-Afrika en spesifiek Mpumalanga word gegee. Die tesis gee 'n oorsig oor brandoorsake, skade wat deur brande veroorsaak word (oppervlakte sowel as finansieHe waarde) en verskeie aspekte rakende brandbestuur soos tyd van ontstaan en reaksietye. Data is volledig ontleed om moontlike verwantskappe te probeer vind tussen vak-, klimaat- en brandbestuursveranderlikes. 'n Oorsaak- en frekwensievoorspellingsmodel is ontwikkel wat brandbestuurders sal help om waarskynlikhede van brande per oorsaak te identifiseer. Statistiese riglyne ten opsigte van bestuursbeplanning vir weerligvure, brande deur heuninguithalers, brande as gevolg van plantasiewerksaamhede en ook brande deur mense (publiek, eie arbeid en kontrakteurs) is daargestel. Brandrekords wat oor 'n periode van 47 jaar gestrek het, is ontleed. Afleidings wat uit die resultate gemaak is, kan benut word om riglyne daar te stel vir strategiese brandbestuur in Mpumalanga Noord plantasies. Die hoof gevolgtrekkings is: • Statistiek van vorige vure is baie nuttig in brandbestuursbeplanning aangesien dit waardevolle inligting verskaf oor brand oorsake, tyd van ontstaan, historiese werkverrigting rakende reaksietye en blustye, sowel as skade wat veroorsaak is. Die gemiddelde oppervlakte beskadig in die studie area is 209.9 ha, of 0.43% van die plantasie oppervlakte per jaar. Menslike aktiwiteite (brandstigting, rakers, piekniekvure, kinders en vure van bure) het die meeste brande veroorsaak (48%), gevolg deur weerlig (22%). Sommige plantasies het swak gevaar en het tot soveel as dubbel die aantal vure per 1 000 ha plantasie gehad in vergelyking met ander plantasies in dieselfde geografiese gebied. Daar is duidelike patrone gevind in die frekwensie van brande per oorsaak oor maande van die jaar. Hierdie patrone is nuttig vir die ontwikkeling van bestuurstrategie vir brande wat veroorsaak word deur heuningversamelaars, weerlig en werkverwante aktiwiteite (plantasieaktiwiteite).

Understanding the quality of the air we breathe is critical in quantifying the impact that atmospheric chemistry has on health. Poor air quality increases the risk of heart and lung diseases as well as having a detrimental effect on climate, ecology and the built environment. The burning of fossil fuels and plant matter (biomass burning) creates large quantities of gases and particulate matter that impact air quality and the air we breathe. Biomass burning is estimated to contribute 400 Tg of non-methane organic compounds, 40 Tg of methane and 7.1 Tg of nitrogen oxides to the atmosphere each year. This thesis aims to better understand the role of biomass burning on air quality and tropospheric chemistry. The in depth analysis presented here addresses of the impact of boreal biomass burning in North America on air quality, in particular, carbon monoxide (CO) and ozone (O3). By using a number of different modelling techniques along with data collected from a field campaign and satellites the transport and chemistry of biomass burning emissions were analysed and quantified. The first research chapter of the thesis used the GEOS-Chem atmospheric chemistry transport model to interpret aircraft measurements of CO in biomass burning outflow taken during the 2011 BORTAS-B campaign over Canada. The model has some skill reproducing the observed variability, but has a positive bias for observations < 100 ppb and a negative bias for observations > 300 ppb. It was found that observed CO variations are largely due to fires over Ontario, with smaller and less variable contributions from fossil fuel combustion from eastern Asia and NE North America. To help interpret observed variations of CO an effective physical age of emissions (¯A) metric was developed. It was found that during BORTAS-B the age of emissions intercepted over Halifax, Nova Scotia is typically 4–11 days, and on occasion as young as two days. The analysis shows that ¯A is typically 1–5 days older than the associated photochemical ages inferred from co-located measurements of different hydrocarbons. It is argued that a robust observed relationship between CO and black carbon aerosol during BORTAS-B (r² > 0.7), form the basis of indirect evidence that aerosols co-emitted with gases during pyrolysis markedly slowed down the plume photochemistry during BORTAS-B with respect to photochemistry at the same latitude and altitude in clear skies. The second research chapter focuses on O3 production downwind from boreal biomass burning. Using the GEOS-Chem model, the O3 chemistry within a biomass burning plume from a fire on 17 July 2011 in mid-Canada was examined. The model shows a significant positive bias (~20 ppb) in reproducingO3 mixing ratios over North America for July 2011 when compared to observations. Reducing NO emissions from lightning and fossil fuel by 50% and 54% respectively reduced this bias to ~10 ppb. The cause of the remaining bias is uncertain. Using a novel technique with the model, the centre of the biomass burning plume was tracked and O3 concentrations and chemistry was extracted from the centre of the plume. The biomass burning enhanced O3 concentrations throughout the plume by between 1 – 20 ppb when compared with the same plume path with no biomass burning. The plume was characterised as being NOx-rich for the initial four days of transport. The sensitivity of the O3 chemistry to different emissions was calculated and it was found that the O3 is initially highly sensitive to NO emissions from biomass burning and then to NO emissions from fossil fuels as it travels across an urban area surrounding Quebec City. The O3 net production was found to initially decrease with an increase in NO but increase further downwind. The final research chapter of the thesis uses long-term satellite observations to evaluate natural variability in CO concentrations over the North Atlantic. 15 years of MOPITT CO column observations were used along with modelled CO from the GEOS-Chem model. The model was evaluated against the MOPITT overpass and shows a negative bias of between -8% and -24% over the northern mid-latitudes with the largest bias seen in spring. The model has a large positive bias (8% – 40%) over the Amazon,West Africa and Indonesia through all seasons. Using Empirical Orthogonal Function (EOF) analysis on the MOPITT and GEOS-Chem CO columns shows the largest mode of variability seen in the North Atlantic to be the oxidation of methane for winter and spring, biomass burning during summer and fossil fuel combustion from East Asia during autumn.

This thesis presents a summary of published data on forest fires and which during these fires are likely to arise and endanger human lives, especially professional firefighters. It also contains general information on fires and forests, but there are also listed the most common causes of this type of fire. It also deals with the knowledge of the public, both professional and lay, of substances generated during forest fires.

Os incêndios estão entre as principais causas da perda de diversidade em florestas tropicais e aparentemente seus impactos são ainda mais intensos em áreas de floresta sob efeito de borda. Com o objetivo de quantificar os danos causados pelo fogo sobre o ecossistema e verificar se a dimensão dos danos e a resiliência do ecossistema estão correlacionados com a distância da borda, foram analisados o banco de sementes e a comunidade vegetal em um trecho de floresta estacional semidecidual, na estação ecológica dos Caetetus, Gália, SP. A área experimental compreende dois setores: a floresta queimada, alvo de incêndio acidental e a floresta não queimada adjacente, utilizada como controle. Cada setor foi subdividido em duas faixas de distância da borda da floresta: 0-20 m e 20-50 m. Para o estudo do banco de sementes foram coletadas amostras nas quatro situações de amostragem, cinco dias após o fogo. Para avaliação dos impactos do incêndio sobre a comunidade vegetal e monitoramento da regeneração dos estratos arbóreo e regenerante na área queimada, foram amostrados cinco transectos de 10 x 50 m sentido borda - interior, avaliados aos seis, 15 e 24 meses após a passagem do fogo. O mesmo desenho amostral foi utilizado na área não queimada, em um único levantamento. Visando verificar se a eliminação de gramíneas invasoras e lianas facilitaria a regeneração da comunidade arbórea, foram instalados cinco transectos adicionais de 10 x 20 m, perpendiculares à borda, nos quais foram efetuadas quatro operações de retirada de lianas e capins, em um período de 24 meses. No banco de sementes, tanto a densidade quanto o número de espécies foram consideravelmente inferiores na área queimada (97 sementes.m-2, de 26 espécies) em comparação com a área não queimada (257 sementes.m2, de 40 espécies). A avaliação dos impactos na estrutura da floresta revelou que o fogo foi mais intenso na faixa mais externa da borda, em que houve perda de 100% da biomassa arbórea, enquanto na faixa mais interna a perda foi de 89%. Em comparação com a floresta não queimada, a comunidade vegetal na área atingida pelo fogo apresentou 43 espécies a menos aos seis meses, diferença que diminuiu para 14 espécies aos 24 meses. A resiliência, analisada com base na recuperação da biomassa arbórea, é maior na faixa mais interna, devido às espécies pioneiras oriundas de sementes que se desenvolvem rapidamente. A rebrota de árvores atingidas pelo fogo também é maior na faixa mais distante da borda e contribuiu significativamente na recuperação da riqueza. O controle de gramíneas e cipós apresentou efeito benéfico exclusivamente para o estrato arbóreo e apenas na faixa de 0-10 m de distância da borda da floresta, proporcionando aumento de área basal, densidade total de plantas e cobertura de copas. Os resultados das operações de manejo indicam que técnicas complementares devem ser aplicadas, visando à facilitação da restauração da floresta após o incêndio. O fogo mostrou-se como elemento de degradação, desde o banco de sementes até o estrato arbóreo. Ainda que a floresta tenha recuperado parte de sua riqueza em dois anos, este processo é lento, caracterizando baixa resiliência, especialmente na faixa mais externa da floresta onde o fogo é ameaça permanente. / Fire is one of the main factors causing biodiversity losses in tropical forests and such losses are reported to be still more intense in forest edges. With the aim of quantifying the damages by fire on the seasonal semideciduous forest, and verifying weather their extension and the ecosystem resilience are correlated to the border distance or not, we analyzed both the seed bank and the plant community after fire in the forest edge of Caetetus ecological station (Gália, São Paulo state, Brazil). The experimental area comprised two sectors: 1) burned area (after an incidental fire in October 2003), and 2) unburned neighboring area (control). Every sector was divided in two strips, according to the distance from the edge: 0-20 m and 20-50 m. The soil seed bank was surveyed in the four situations described, five days after fire. Plant community structure and regeneration were assessed in five permanent transections (10 x 50 m), instaled from the edge to the interior, where plant individuals were measured and identified, at six, 15, 18 and 24 months after fire. The same design was repeated once in the unburned area, for comparison. Additionally, with the aim of verifying the hypothesis that arboreal community regeneration after fire could be improved by controlling lianas and invasive grasses, five managed transections (10 x 20 m each) were installed, perpendicular to the edge, where grasses and lianas were eliminated four times within a 24 months period. Fire effects on the ecosystem were remarkable. Richness and density of the seed bank after fire (97 seeds.m-2, 26 species) were considerably lower than in the neighboring unburned forest (257 seeds.m-2, 40 species), the difference being still larger close to the edge (0-20 m). Forest structure was totally changed, 100% of the arboreal biomass lost in the external strip (0-20 m) and 89% lost in the internal strip (20-50 m). Six months after fire the burned area had 43 species, less than the unburned forest. Richness has been slowly recovered and, 24 months after fire, that difference had decreased to 14 species. Resilience, analyzed in terms of biomass recovery, is higher in the internal strip (20-50 m from the edge), where pioneer species quickly regenerate from seeds. Sprouting, which has also been more effective in the internal strip, has equally contributed to the richness recovery. Eliminating grasses and lianas favored only arboreal species in the external strip (0-10 m) where their density, basal area and crown cover were slightly higher than in the unmanaged plots. We consider that this technique, alone, can not be recommended to improve the forest restoration after fire. Fire was, though, a very strong degrading agent in the studied forest, almost completely destroying the arboreal biomass, besides the remarkable reduction in species richness in the seed bank and in the arboreal layer as well. Even though the forest richness and biodiversity have been recovered, that has been a very slow process (low resilience) specially close to the forest edge, where fire is a permanent threat.

Neste trabalho o aerossol atmosférico, presente em diferentes localidades da região Amazônica, foi caracterizado. Foram estudadas as emissões biogênicas naturais da floresta, bem como as emissões de material particulado para a atmosfera, devido à ocorrência em larga escala de queimadas na Amazônia durante a estação seca. O aerossol atmosférico foi coletado durante um período de aproximadamente três anos em duas diferentes regiões da Amazônia, utilizando-se amostradores de particulado fino e grosso. As amostragens foram realizadas na região de Alta Floresta (MT) e de Serra do Navio (AP). Essas regiões são representativas de condições atmosféricas distintas: as emissões de queimadas são pouco significativas para a composição do aerossol atmosférico de Serra do Navio, dominada pela emissão biogênica natural da floresta. A atmosfera de Alta Floresta caracteriza as regiões da Amazônia onde o impacto das queimadas é expressivo durante a estação seca. Em Alta Floresta foram realizadas amostragens em casas de queima de ouro, caracterizando as emissões de mercúrio para o particulado atmosférico relacionadas com as atividades garimpeiras na Amazônia. Aeronaves foram utilizadas para a amostragem de aerossóis durante as estações secas de 1992 e 1993, caracterizando as concentrações atmosféricas dos aerossóis de queimadas ao longo de vastas regiões da Amazônia. As amostras foram analisadas utilizando-se diversas técnicas analíticas nucleares: \"Particle lnduced X-ray Emission\", para a quantificação dos elementos traços com número atômico superior a 11; \"Induced Gamma-ray Emission\", para a quantificação de Na; \"Scanning Proton Microprobe\", para a caracterização de partículas individuais de aerossol. A técnica de reflectância foi utilizada na quantificação de \"black carbon\" e a concentração total de aerossóis atmosféricos foi determinada por análise gravimétrica. \"Cold Vapor Atomic Absorption Spectroscopy\" foi utilizada na quantificação de mercúrio presente no particulado atmosférico nas casas de queima de ouro de Alta Floresta. Para as amostras de particulado fino de Serra do Navio, constituintes iônicos do aerossol foram quantificados utilizando-se cromatografia iônica. Técnicas estatísticas multivariadas foram utilizadas na identificação e na caracterização das fontes de aerossol atmosférico presentes nas diferentes regiões. Observaram-se elevadas concentrações de aerossóis atmosféricos em vastas regiões da Amazônia devido às emissões de queimadas para o período compreendido entre os meses de junho e setembro. As emissões de queimadas são predominantes na fração fina do aerossol atmosférico, sendo caracterizadas por grandes emissões de \"black carbon\", S e K. As elevadas emissões de aerossóis pelas queimadas estão correlacionadas com o aumento do \"Aerosol Optical Thickness\" da atmosfera durante a estação seca na Amazônia. O aerossol de Serra do Navio é caracterizado por emissões biogênicas, apresentando forte influência marinha. Foi observada uma semelhança entre as composições do aerossol de origem biogênica presentes em Serra do Navio e em Alta Floresta. A presença de elementos traços característicos de partículas de solo associados à contribuição marinha para o aerossol de Serra do Navio indica o transporte de aerossóis provenientes da África para a América do Sul. Observaram-se elevadas concentrações de mercúrio associadas ao aerossol atmosférico presente no interior das casas de queima de ouro de Alta Floresta. A concentração de mercúrio, associada ao particulado atmosférico, representa cerca de 3% da concentração total de mercúrio no interior das casas de queima. Uma fração substancial do mercúrio está associada ao particulado fino, facilitando o seu transporte juntamente com o aerossol emitido nas queimadas. A fração não volátil do mercúrio associado ao aerossol é considerável, estando fortemente ligada às partículas em suspensão na atmosfera. / This work presents the characterization of the atmospheric aerosol collected in different places of the Amazon Basin. We studied both the biogenic emission from the forest and the particulate material which is emitted to the atmosphere due to the large scale man-made burnings during the dry season. The samples were collected during a three year period at two different locations in the Amazon, namely the Alta Floresta (MT) and Serra do Navio (AP) regions, using stacked unit filters. These regions represent two different atmospheric compositions : the aerosol is dominated by the forest natural biogenic emission at Serra do Navio, while at Alta Floresta it presents an important contribution from the man-made burnings during the dry season. At Alta Floresta we took samples in gold shops in order to characterize mercury emission to the atmosphere related to the gold prospection activity in Amazon. Airplanes were used for aerosol sampling during the 1992 and 1993 dry seasons to characterize the atmospheric aerosol contents from man-made burnings in large amazonian areas. The samples were analyzed using several nuclear analytic techniques: Particle Induced X-ray Emission for the quantitative analysis of trace elements with atomic number above 11; Particle Induced Gamma-ray Emission for the quantitative analysis of Na; and Proton Microprobe was used for the characterization of individual particles of the aerosol. Reflectancy technique was used in the black carbon quantification, gravimetric analysis to determine the total atmospheric aerosol concentration and Cold Vapor Atomic Absorption Spectroscopy for quantitative analysis of mercury in the particulate from the Alta Floresta gold shops. Ionic cromatography was used to quantify ionic contents of aerosols from the fine mode particulate samples from Serra do Navio. Multivariated statistical analysis was used in order to identify and characterize the sources of the atmospheric aerosol present in the sampled regions. High atmospheric aerosol concentrations were observed in large amazonian areas due to emission from man-made burnings in the June-September period. The emission from burnings dominate the fine mode of the atmospheric aerosol with characteristic high contents of black carbon, S and K. Aerosols from biomass burnings are correlated to the increase in the Aerosol Optical Thickness of the atmosphere during the Amazonian dry season. The Serra do Navio aerosol is characterized by biogenic emissions with strong marine influence. The presence of trace elements characteristic of soil particulate associated with this mar\'ine contribution indicates the existence of aerosol transport from Africa to South America. We observed common composition characteristics in the biogenic emission aerosols from Serra do Navio and Alta Floresta. High mercury concentrations associated to the atmospheric aerosol inside the Alta Floresta gold shops were observed. About 3% of the total mercury concentration inside the gold shops is associated to the atmospheric particulate. A substantial part of this mercury is associated to the fine mode, easing its transportation with the biomass burnings aerosol. A large fraction of the mercury associated to the aerosol is non-volatile, being strongly tied to the particulate in suspension in the atmosphere.

Este trabalho trata do estudo do transporte atmosférico de emissões de queimadas na região Amazônica e Centro-Oeste do Brasil. Em consequência da queima de biomasa, gases (CO IND.2, CO, CH IND.4, etc.) e partículas são emitidas para a atmosfera, os quais podem provocar mudanças no equilíbrio climático e biogeo químico do planeta Terra, em várias escalas. O estudo é realizado por meio de simulação numérica dos movimentos da atmosfera utilizando o modelo atmosférico RAMS (Regional Atmospheric Modeling System). São utilizados os métodos Lagrangiano e Euleriano. No método Langrangiano o transporte é estudado por meio do cálculo de trajetórias cinemáticas 3d de massas de ar, utilizando o campo de vento simulado, na escala resolvida. A posição inicial das massas de ar é obtida a partir de mapas de queimadas detectadas por sensores a bordo de satélites como o AVHRR da série NOAA. Uma metodologia simples que contabiliza os efeitos de processos convectivos úmidos da escala sub-grade na velocidade vertical da parcela de ar para modelo atmosféricos de baixa resolução, é introduzida no modelo de trajetórias. No método Euleriano, é resolvida a equação de conservação de massa dos principais elementos emitidos. Para tanto é introduzido um modelo de fontes emissoras de gases/partículas associadas à queimadas em floresta tropical e cerrado da América do Sul, distribuídas espacial e temporalmente através da assimilação diária de mapas de posição de queimadas produzidos por sensoriamento remoto. Os termos de advecção na escala resolvida e o transporte turbulento na escala sub-grade são resolvidos utilizando as parametrizações próprias do modelo RAMS. É introduzida uma parametrização do transporte sub-grade associado às circulações úmidas e profundas não resolvidas explicitamente pelo modelo, devido a baixa resolução espacial. Termos sumidouros associados a processos genéricos de remoção/transformação de gases/partículas são ) também parametrizadas e inseridos na equação de conservação de massa. A metodologia é aplicada a um estudo de caso ocorrido no mês de julho de 1993. São modelados o transporte turbulento na camada limite planetária, os efeitos transientes associados à entrada de frentes frias vindo do sul do continente e à convecção na bacia Amazônica no transporte convectivo de poluentes, bem como o transporte associado aos sistemas de larga escala. São observados padrões de exportação continental de poluentes, com saídas ocorrendo a noroeste da América do Sul em direção ao oceano Pacífico, e a sudeste em direção ao oceano Atlântico. / A study about the atmospheric transport f biomass burning emissions in the Amazon and the Central of Brazil is presented. Gases (CO IND.2, CO, CH IND.4, etc.) and particles emitted to the atmosphere due the biomass burning are responsible for the climatic and biogeochemical budget changes in the Earth planet, in many scales. This study is carried out through a numerical simulation of the atmospheric motions using the atmospheric models RAMS (Regional Atmospheric Modeling System). Lagragian and Eulerian methods are used. In the Lagrangian method the transport is studied through 3D kinematic air mass trajectories calculation, using simulated wind fields, in a resolved scale. The initial position of the air masses is obtained from biomass burning spots maps, derived from satellite sensors (AVHRR from NOAA series). A simple methodology to take into account the sub-grid effects of wet convective process in the vertical velocity of the air parcels, for low resolution atmospheric models is introduced in the trajectory model. In the Eulerian method the mass conservation equation is resolved for the main elements emitted. A model of gases and particles sources emissions is introduced associated with biomass burning in South America tropical forest and savanna, spatially and temporally distributed and daily assimilated, according to the biomass burning spots defined by remote sensing. The advection in a resolved scale and turbulent transport, in a sub-grid parameterization associated to wet and deep circulation not explicitly resolved by the model due its low spatial resolution is introduced. Sinks associated with generic process of removal/transformation of gases/particles are parameterized and introduced in the mass conservation equation. The methodology is applied to a case study on July 1993. The turbulent transport in the planetary boundary layer the transient effects in the convective transport of pollutants associated with cold fronts from the south and convection in the Amazon basin and the transport associated with the large scale systems are modeled. Patterns of pollutant exportation are observed with a corridor to the Pacific Ocean in the South America Northwest and another in the Southeast to the Atlantic Ocean.

Fire ecology encompasses the study of the factors, biotic and abiotic, that influence the occurrence of fire in an area, as well as the effects fire has on the flora and fauna native and non-native to the region (Whelan 1995). Fire has had a major influence on shaping biomes as we see them today. Fire has had an effect on vegetation much before the evolution of Homo on Earth (Keeley and Rundel 2005, Pausas and Keeley 2009, Midgley and Bond 2011). With the evolution and expansion of Homo across Earth, fire has been tamed, and then generated and used over time to yield landscapes that were suitable for their existence (Pyne 1991, Bowman et al. 2009, Archibald et al. 2012). Thus, fire, vegetation and humans were, and still are, inextricably linked in certain biomes on Earth. The best examples are observed in tropical savannas and grasslands, biomes that experience distinct seasonality in climate and are thus prone to frequent fires caused either by lightning or by humans (Keeley and Rundel 2005, Archibald et al. 2012). At the other end of the spectrum of tropical vegetation types are rainforests where the occurrence of fires is constrained by a perpetually moist environment (Meyn et al. 2007, van der Werf et al. 2008), in the absence of manipulation of the forest landscape by humans. Frequent fires have been documented to alter structure and cause a decline in forest diversity in rainforests (Cochrane and Schulze 1999, Cochrane 2003), whereas fire exclusion in mesic savannas leads to increases in biomass and transition to forest ecosystems (Bond et al. 2003, Bond et al. 2005 and references therein). A tropical biome that lies between these two extremes of vegetation types is the Seasonally Dry Tropical Forest (SDTF) where the occurrence of fire is common, but for which there are contrasting views on the effect of fire on this system (Saha and Howe 2003, Otterstrom et al. 2006 as examples). Current forest management policies in SDTF areas, especially in India, actively aim to exclude fire from these forests mostly because of the perception held by forest managers and the general public that fire has negative effects on forests. However, very few scientific studies have explored the ecology of fire in SDTFs. In order to formulate fire management policies, it is necessary to have a more comprehensive understanding of the ecology of fire in this tropical forest type. This thesis addresses two components of fire ecology as applied to SDTFs. The first is how fire is influenced by the environment, and the second, how fires influence the biotic community particular to SDTFs. The study was carried out in an SDTF in southern India where fire is a common occurrence -the forests of Mudumalai – a protected area that exhibits a range of SDTF vegetation types, from moist deciduous to dry thorn forest, corresponding to a rainfall gradient. Fire influenced by the environment: For this section, the influence of fuel load, fuel moisture and ambient weather on area burnt, fire occurrence and fire temperatures were studied in the SDTF vegetation types of Mudumalai. The extent of fire (area burnt) in an ecosystem differs according to the relative contribution of fuel load and fuel moisture available (Meyn et al. 2007). At a global scale, these factors vary along a spatial gradient of climatic conditions and are thus “varying constraints” (Krawchuk and Moritz 2011) on fire activity in natural ecosystems (Meyn et al. 2007, Krawchuk and Moritz 2011). Moist ecosystems such as tropical rainforests are at one end of the spectrum where fire activity is constrained by fuel moisture. At the other end are arid ecosystems, such as deserts, where fire activity is limited by the presence of fuels. The potential for the globally widespread seasonally dry tropical forests (SDTFs) to be placed as a single entity in this framework was examined by analyzing the interacting effects of fuel load and fuel moisture on the extent of fire in Mudumalai. Logistic regression was used to model proportion area burnt in a given year with factors that would influence fuel load and fuel moisture – these were proportion area burnt the previous year, wet season rainfall the previous year and early dry season rainfall. Modelling was conducted at two levels – the overall landscape and within four defined moisture regimes (between 700 and 1700 mm yr-1) – using a dataset of area burnt and seasonal rainfall from 1990 to 2010. The landscape scale model showed that the extent of fire in a given year within this SDTF is dependent on the combined interaction of seasonal rainfall and extent burnt the previous year. However, within individual moisture regimes the relative contribution of these factors to the annual extent burnt varied – early dry season rainfall (i.e. a moderator of fuel moisture) was the predominant factor in the wettest regime, while the previous year’s wet season rainfall (i.e. a proxy for fuel load) had a large influence on fire extent in the driest regime. Thus, the diverse structural vegetation types associated with SDTFs across a wide range of rainfall regimes would have to be examined at finer regional or local scales to understand the specific environmental drivers of fire. While the extent burnt in SDTFs is largely dependent on climatic influences, the probability of ignition has not been characterized for SDTFs. Anthropogenic fires are a regular occurrence during the dry season in SDTFs (Stott et al. 1990). We investigated if the occurrences of anthropogenic fire in Mudumalai were associated with any particular weather conditions during the dry season. Logistic regression between probability of a fire day and weather variables -seasonal rainfall, ambient relative humidity and temperature -was examined during the dry seasons of 20042010 in Mudumalai. Fire incidence data was obtained from the Fire Information for Resource Management System (FIRMS; NASA 2002) and weather data from two automatic weather stations within Mudumalai. The analysis showed that days with high probabilities of fire occurrence were associated with low levels of early dry season rainfall, low daily average relative humidity, and high daily average temperatures. These weather conditions are known to influence moisture levels of fine fuels (Viney 1991, Archibald et al. 2009). In Mudumalai as well as other SDTFs the primary fuels for fires are fine fuels such as litter and dried grass that accumulate on the forest floor during the dry season. Our results suggest that the occurrence of fire is moderated by environmental conditions that reduce or enhance the flammability of fine fuels in the dry tropics. A quantitative framework for assessing risk of a fire day has been proposed as an outcome of this analysis to assist forest managers in anticipating fire occurrences in this SDTF, and possibly for those across south Asia. Of the various components of a fire regime, fire intensity is an important aspect. High fire temperatures (one measure of fire intensity, Keeley 2009) and resulting soil temperatures would have an effect on soil properties as well as plant species demography and community structure (Moreno and Oechel 1991, Neary et al. 1999, Morrison 2002). Fires that occur frequently in a region could vary in their intensity and severity depending upon the amount of fuel available and ambient weather conditions (Stinson and Wright 1969, Stott 1986, Stronach and McNaughton 1989, Ansley et al. 1998, Wotton et al. 2012). However, this relationship has not been examined in a multiple regression framework for SDTFs. Fire temperature was recorded and its relationship with ambient weather and fuel load was studied in two SDTFs of southern India -Mudumalai and Biligiri Rangaswamy Temple (BRT) Wildlife Sanctuary in Karnataka. During “controlled burns” conducted by the forest department staff in these reserves in February and March 2010, temperature indicating lacquers on mica sheets were used to measure fire temperature at several points at ground level and one cm below the ground. Biomass was harvested close to the temperature measurement points to estimate fuel load and fuel moisture. Ambient weather conditions were recorded during the controlled burn when the flame passed over the indicators. Temperatures recorded at ground level ranged from <79oC to 760oC, with the most frequently recorded temperatures between 343-399 oC and 510566 oC. Temperatures measured one cm below the ground ranged from <79oC to 302oC, with a majority of the indicators recording temperatures in the <79oC category. Ground-level temperatures increased with increasing biomass. A linear regression of ground-level temperatures with fuel load and ambient weather conditions of relative humidity and temperature was found to explain most of the variation in the data. Ground-level fire temperatures increased with increasing fuel load, but were also found to be lower at higher relative humidities at a given temperature. In order to reduce the intensity of forest fires that occur accidentally during the dry season, we recommend that fuel loads be reduced in the forest by prescribed burning early in the dry season. This applies especially to areas where there is accumulation of biomass over years, such as that of the tall grass Themeda cymbaria found predominantly in dry deciduous forest types. If prescribed burning is incorporated in fire management policies for these forests, then the season of burning will be important to consider. It is known from ecosystems where prescribed burning is regularly applied that early dry season fires are less intense than late dry season fires (Williams et al. 1998). However, this has not been systematically investigated for SDTFs. Through a burning experiment carried out in private land with vegetation type similar to tropical dry thorn forest, we investigated differences in area burnt, ground-level fire temperatures and soil temperatures one cm below the ground in the early dry season in January, late dry season in April and and early wet season in June. We also examined differences in fuel load, fuel moisture, soil moisture and weather conditions of ambient relative humidity (RH), temperature and wind speed in these phases; these factors could be responsible for observed differences in fire and soil temperatures or area burnt. Although area burnt was not significantly different between the early and late phases of the dry season, fire and soil temperatures were significantly lower in the former. The late dry season was characterized by distinctly higher fuel loads, lower fuel moisture, lower relative humidity, higher ambient temperatures and higher wind speeds compared that measured in the early dry season. Differences in soil temperature between these months may be attributed to the increase in fuel load since there were no significant differences in soil moisture. Fire spread was limited in the experimental plots in the early wet season in June, probably due to significantly higher levels of fuel moisture in this month; the resultant fire and soil temperatures recorded were low. Forest management should, therefore, consider early dry season burns in the month of January for prescribed burns in the sanctuary, although this would have to be tested in other SDTF vegetation types with more variable fuel load, fuel moisture and weather conditions. Fire’s influence on the biotic community: Concerns regarding the regeneration capacity of woody species in SDTFs have been voiced with respect to increasing frequencies of fire (Saha and Howe 2006, Kodandapani et al. 2008). Fire is known to cause high mortality of individuals of small size (Swaine et al. 1990, Suresh et al. 2010). However, mortality has been examined for large size classes, and not for seedlings. It is essential to understand the dynamics of seedlings and their contribution to the regeneration potential of SDTFs. Woody species in SDTFs are known to have traits that help them recover from recurring disturbances, such as sprouting from underground root stocks (Vieira and Scariot 2006). Another trait may relate to growth rates of seedlings. Growth rates of seedlings (defined in this study as established individuals between 10 and 100cm height) after dry season (February-March) fires were compared between adjacent pairs of burnt and unburnt transects established at eight sites in Mudumalai across vegetation types of moist deciduous, dry deciduous and dry thorn forest. The growth of grasses, a possible competitor for resources, was also monitored at each site. Seedling and grass heights were monitored at 3-month intervals between August 2009 and August 2010. A second fire in March 2010 affected transects at two sites in Mudumalai. Seedling and grass heights were monitored for two enumerations till August 2010 subsequent to the second fire at these two sites. A total of 1032 individuals across 58 woody species were enumerated. High seedling survivorship (>95%) was observed in both burnt and unburnt areas. Although seedling heights were significantly different between burnt and unburnt areas at the start of the enumeration in August 2009, heights were comparable within a year and a half of the fire. Comparable seedling heights in such a short time span were because of distinctly higher growth rates of seedlings in burnt areas compared to unburnt areas after the fire event, particularly during the pre-monsoon season. Grass biomass (volume), on the other hand, was significantly different between burnt and unburnt areas at both the first and last enumerations. Grass growth (change in volume) did not differ between burnt and unburnt areas. Rapid growth by seedlings after a fire implies adaptation through the use of stored resources for growth, possibly aided by lower competition from grasses, in order to attain a certain size before the subsequent return of unfavourable factors such as a recurrent fire event. Conclusions: The results from the study point to climatically driven fire regimes in an SDTF in southern India, with daily influences of weather conditions during the dry season on fire occurrences. Fire intensities increase with increasing fuel loads in these forests, moderated by weather conditions such as RH and temperature. Since fires are an anthropogenic phenomenon in these forests, active management with the use of prescribed fires in the early phase of the dry season is a possible option to control late dry season fires that would be higher in intensity. The current woody tree species assemblage in this southern Indian SDTF is resilient to fires at the seedling stage, with established individuals exhibiting high survivorship and rapid growth after a fire. However, the effects of fires of varying intensities on the regenerative capacity of the seedlings are not known. The effect of fire on habitat utilisation by large herbivores, or the impact of fire on the faunal community in general has not been studied for vegetation types that comprise SDTFs. The effect of fire exclusion on the ecology of SDTFs will provide useful information that can feed into management policies for this ecosystem type. These are potential areas of research for the future. Fire, if managed wisely, can be an effective tool for the conservation of SDTFs across south and southeast Asia.

The petroleum industry has one of the most complex supply chains in the world. A unique characteristic of Petroleum Supply Chain (PSC) is the high degree of uncertainty which propagates through the network. Therefore, it is necessary to develop quantitative models aiming at optimizing the network and managing logistics operations. This work proposes a deterministic Mixed Integer Linear Program (MILP) model for downstream PSC to determine the optimal distribution center (DC) locations, capacities, transportation modes, and transfer volumes. Three products are considered in this study: gasoline, diesel, and jet fuel. The model minimizes multi-echelon multi-product cost along the refineries, distribution centers, transportation modes and demand nodes. The relationship between strategic planning and multimodal transportation is further elucidated. Furthermore, this work proposes a two stage Stochastic Mixed Integer Linear Program (SMILP) models with recourse for PSC under the risk of random disruptions, and a two stage Stochastic Linear Program (SLP) model with recourse under the risk of anticipated disruptions, namely hurricanes. Two separate types of mitigation strategies – proactive and reactive – are proposed in each model based on the type of disruption. The SMILP model determines optimal DC locations and capacities in the first stage and utilizes multimode transportation as the reactive mitigation strategy in the second stage to allocate transfer volumes. The SLP model uses proactive mitigation strategies in the first stage and employs multimode transportation as the reactive mitigation strategy. The goal of both stochastic models is to minimize the expected total supply chain costs under uncertainty. The proposed models are tested with real data from two sections of the U.S. petroleum industry, PADD 3 and PADD 1, and transportation networks within Geographic Information System (GIS). It involves supply at the existing refineries, proposed DCs and demand nodes. GIS is used to analyze spatial data and to map refineries, DCs and demand nodes to visualize the process. Sensitivity analysis is conducted to asses supply chain performance in response to changes in key parameters of proposed models to provide insights on PSC decisions, and to demonstrate the impact of key parameters on PSC decisions and total cost. / Upper Great Plains Transportation Institute (UGPTI) / Mountain Plains Consortium (MPC)