OEFS Abstracts - Spring 2009

Jacob Berv - Investigation of the Mountain Pine Beetle Infestation in British Columbia, Canada

Warming climate trends have stimulated the current infestation of Dendroctonus ponderosae in the pine forests of British Columbia to epidemic proportions. A brief review of the beetle’s ecology sheds light on remote sensing opportunities, which have been established as effective methods of monitoring the “red-stage” outbreak stage from space. Using 250m MODIS Vegetation Indies and a 90m USGS DEM, comparative EVI analysis demonstrates two hypotheses which are testable using unsupervised classifications; that the area of damaged forest has not remained constant over the last eight years, and that the spread of the infestation has been influenced by the topology of the region. Accuracy assessment with ground truthing data or established forestry data is necessary to support the method described in this paper. While researchers work for an ecological solution, remote sensing is playing a critical role in monitoring the situation from space.

 

Ian Cummins - Deforestation in Santa Barbara Honduras

The patterns and processes of deforestation and land use conversion are analyzed within the North-Central region of Honduras. Time-Step ASTER images are used to determine both forest cover and rates of land use change between 2003 and 20058. A LANDSAT 5 TM covering parts of the region is analyzed using a number of classification types and classification techniques. A series of unsupervised classifications with varying numbers of iteration and classification methods were optimized to distinguish forested lands from other land cover categories. Digital Elevation Models were also utilized to determine the spatial distribution of forested land. A supervised classification method utilizing geo-referenced photographs waypoints was also utilized used to classify land cover type, these results are compared to what was gained by the unsupervised classification. Overall, both classification schemes satisfactorily identified forest cover as well as other land cover types. Forest cover in the area was found to have decreased significantly between 2003 and 2005. Forested land was also found to be highly correlated with increasing topographical relief with sharply slope uplands tending to be heavily forested while higher quality lowlands area were nearly universally dedicated to agriculture and pasture land. Analysis of the region indicates that despite low rates of deforestation, forest cover is patchy, isolated and highly degraded and that reforestation efforts should be undertaken.

 

Christine Trac - Forest Change in China: are China’s Forestry Statistics Supported by Remote Sensing Data?

After a century of intensive forest exploitation and destruction the Chinese government has made ambitious efforts to increase forest cover across the country, particularly as an effort to control soil erosion and prevent flooding. Forest statistics reveal that China’s forests are rapidly expanding through afforestation and reforestation programs implemented nationwide, however, what these numbers reveal is not straightforward. Changes to China’s forestry accounting have enhanced these numbers while past forestry practices hide the quality of forests. To understand how China’s forests are changing it is necessary to have another measure. Remote sensing can serve as an alternative to official measures and provide a metric that subjects forests uniformly across time. I am interested in exploring the change in China’s forests in terms of change in area and change in quality. My study will use remote sensing data to examine forest change in Yanyuan Country, China. Based on my knowledge of the forests in this region, I hypothesize that remote sensing data will reveal there has been minimal forest change in this area. I also hypothesize that the quality of these forests has decreased over time – a result of afforestation efforts occurring at near equivalent rates of deforestation.

 

Peter Coutros - The Timbuktu Expedition Project

Several dramatic climate oscillations plagued the native inhabitants of this region, forcing them to mold their societies around their conditions. This study will hopefully shed some light onto the timing of these oscillations, and in comparison with the archaeological record, be the basis for a chronology of significant change in the sociopolitical structure of the ancient Malians. However, in order to do this, more excavations will need to be carried out in the region. To make this more temporally, and financially practical, remote sensing must be utilized in order to identify sites amid the immense and unforgiving terrain.

 

Lauren Adams - Detecting Water Use in the Desert

The objective of this research is to analyze spatio-temporal human and biogeophysical responses to drought in a trans-boundary arid environment. Specifically, several comparative and mixed methods helped to cluster and label similar land-use and land-cover (LULC) classes, to identify the presence of water and soil moisture content within each land-class within image and to depict land-use patterns during drought conditions within a matrix of desert-agriculture-urban land-uses and a gradient of different bi-national water policies. The primary means by which this analysis was executed was through comparison of several different types of change detection methods pivoting around detection of wetness throughout the landscape. Geographically, the region studied for this analysis spans the Upper Conchos River and the Lower Bravo River Basins near the cities of Ojinaga, Chihuahua, MX and Presidio, Texas. Landsat TM and Landsat ETM imagery for the dates June 14, 1992 and June 26, 2008 were used for the analysis; 1992 is the year the extraordinary drought began in the region, and 2008 denotes the most recent display of adaptation efforts. To observe the water-related changes in the landscape, the Wetness Index is used as the primary means of comparison, although several other methods of comparison and LULC classifications were employed to visually depict and quantify changes in the presence of water and LULC types across space and time, including changes in Brightness, Wetness and Greenness derived from Tasseled Cap Transformations, Albedo, Thermal Infrared Bands, Comparative NDVI images, and various means of image subtraction, classification and re-classification. An obvious shift in the location of wet areas occurred between 1992 and 2008; it was visible in each of the analyses that the climate was drier and cooler in 1992 than 2008. Deeper and sparser wet regions were visible in 1992 whereas in 2008 there average wetness of an area was less, but there were more concentrated areas of saturated soils. Within the Mexico floodplain of the Rio Conchos and the inflow of the Rio Bravo into the confluence of the Conchos and the Bravo the area of saturated soils and surface water increased, whereas downstream of the confluence, particularly on the United States side, where agricultural lands almost disappear entirely, wetness decreases. The conclusion, therefore, of this study was that there is evidence of a socio-economically induced drought within the region, in which wetness increases disproportionately upstream, in Mexico and decreases disproportionately downstream, in the United States, independent of any evidence in support of a hydrologic or meteorological drought found between the years 1992 and 2008.

 

Dana Graef - Land Use Change Around Havana, Cuba 1985-2000

In this paper, I examine land cover change in an area incorporating Havana, Cuba between 1985 and 2000, a period of dynamic social and ecological change before and after the collapse of the Soviet Union. I acquired two Landsat images (Path 16 Row 44) from 1/25/1985 (Landsat TM) and 1/11/2000 (Landsat ETM+) and imported the image files into ERMapper. Prior to classification, I subset the images to ~12,000 km2 and conducted a visual inspection of change with multiple band combinations, and assessed changes in NDVI. I then performed several iterations of supervised classifications to identify developed areas, active fields, fallow fields/bare soil, water, natural vegetation/forests, and a tan/mottled class. I also performed an unsupervised classification. I calculated area summary reports, and selected classes of interest with the equal to class number test to visually depict change. Across the images, I found that developed and agricultural areas diminished, while natural vegetation and the tan/mottled class increased. Similar trends were observed in a post-classification subset of the area around Havana, though the overall NDVI in this area appeared to increase. Finally, I found a significant augmentation of inland water, primarily through increased and expanded reservoirs. Inland water across the image doubled from 1985-2000. In a 618 km2 post-classification subset region in the southwest of the image, water cover quadrupled, and several new hydroelectric projects were built. A methodological comparison of supervised classification, unsupervised classification, and manual digitization of the reservoirs in this region showed a similar rate of increase across the methods, but the supervised classification was more conservative in assigning the water class, and consequently led to lower area summaries of water than the unsupervised and manual methods.

 

Christa Anderson - Land Use Change on Kuku Ranch, Kajiado District, Kenya

This project aims to detect land use change over time on Kuku Group Ranch in Southern Kenya. Much of the land in Kenya is allotted and owned in “group ranches.” Tribal Councils are charged with managing the land within a group ranch for the community residing within the ranch boundaries. Kuku Group Ranch is a 300,000 acre ranch belonging to a Maasai community of 7,000 pastoralist occupants.

Kuku Group Ranch is also home to a local community-based nonprofit organization called Maasai Wilderness Conservation Trust. This nonprofit seeks to aid the local community in the areas of conservation, education, and health. This nonprofit and the community it serves have two primary environmental concerns regarding land use change: 1) agricultural production is expanding rapidly in a fragile ecosystem that has historically been used for herding rather than for agriculture, and 2) illegal timber harvest in the hill regions is increasing and changing the hydrology of the ranch.

This remote sensing project deploys remote sensing methods that attempt to examine both the agricultural and forestry related land use change concerns of Kuku Ranch. In addition to providing useful and desired information for the Maasai Wilderness Conservation Trust about land use change, this project simultaneously allowed the author to gain greater skills in the use of remote sensing methods and techniques.

 

Jacob Munger - Land Use in Bangalore, India

Remote sensing imagery has great potential to assist researchers focused on urban environments. Information obtained from satellite images can help to better understand land use change caused by urban expansion as well as cities’ impacts upon climate. In this paper, I address both of these topics, in the context of Bangalore, India. The first part of the paper discusses comparisons between the temperature in the urban core of Bangalore compared to the temperature in the area immediately surrounding the urban core. The second part of the paper discusses different methods for using satellite imagery to classify land cover in and around Bangalore. The final part of the paper looks at methods for studying land cover change over time in and around Bangalore. All of the analysis for this project was completed using Landsat TM satellite images.

 

Esteban Rossi - Land Cover Change in the Tamed Frontier: Iquitos, Peruvian Amazon

This study describes the land use-land cover dynamics around the city of Iquitos, using Landsat imagery, public land cover maps published by the Peruvian government and information from people from the area. This Peruvian city is interesting because of its biodiversity and because of its geographical isolation. Iquitos is not connected to any road network in Peru or any other Amazonian country.

 

Cica Viana - Land Cover Changes in a Region of Eastern Brazilian Amazon

This study aims to assess urban expansion and land cover change in the Brazilian Amazon. First, different techniques were applied to two MODIS 8-days composite images from 2001 and 2008. However, pixel’s DN values were problematic as a result of the compositing process. Next analysis was using two Landsat images for the city of Porto Velho for years 2000 and 2006, and again various classification methods were applied. The major challenge was to differentiate soil and pasture from urban covers, due to the similarity of their spectral signatures. The Tasseled Cap method allowed determining more uniform training classes, and the supervised classification ran with those classes gave more accurate than the others methods employed. After classifying both images, a change matrix identified transition classes. Major change in land cover was from forest to agriculture/pasture.

 

Jennifer Hoyle - Exploring MODIS Snow Cover Products

The goal of this project is to explore the MODIS Snow Cover Area Product suite, specifically the 500-meter resolution swath product, MODIS/Terra Snow Cover 5-Min L2 Swath 500 m (MOD10_L2) and the MODIS/Terra Snow Cover Daily L3 Global 500m SIN Grid (MOD10A1). Remote Sensing image analysis was applied to an image including New England, the Great Lakes and Canada north to Hudson Bay in order to compare these two products, as well as to re-create the algorithms used to develop the products from reflectance data. Techniques used include image pre-processing, unsupervised classification, subsetting, and thresholding. It was determined that the products are readily available, consistent, and that certain aspects can be re-created with basic knowledge of remote sensing.

 

Marissa Matsler - Calculating Index Variables Using Remote Sensing Techniques

Intra-city variation and change of environmental and socioeconomic factors is often monitored using an integrated approach of remote sensing and GIS techniques. The objective of this project was to calculate two environmental variables; percent greenness and percent impervious surface  utilizing remote sensing techniques in ERMapper 7.2.2. The city of Baltimore was chosen as study site in order to observe, not a growing city, but one in the midst of rigorous redevelopment after a period of decline. Tasseled Cap and NDVI transformations along with Unsupervised Classification were completed to create raster layers that were then quantified by city neighborhood in ArcGIS.

 

Bjorn Fredrickson - Developing a Tea Index for Upper Assam

The goal of this project is to determine the spectral signature of tea, or to develop a “tea index”, in order to track changes in the character of the tea industry in Northeast India. More specifically, I am interested in this project’s potential applications to questions associated with a proliferation of small-scale tea cultivation in Upper Assam over the past two decades; during this period small tea gardens have begun to crop up in the margins between the colonial era tea plantations that cover some 5 percent of the State’s total land area. The Indian State has little sense of how many small-scale growers have come into production, nor how much land is now being cultivated as tea; at a basic level, the ability to identify tea and better track changes associated with land use would prove valuable to the Tea Board of India, Tocklai Research Station (an agricultural research station devoted to agricultural work on tea) and Assam Agricultural University.

 

Xin Zhang - Detecting Landcover Type and Surface Flux Through Remote Sensing in Minn

By detecting the reflectance and radiation from the earth surface in different wavelengths, remote sensing technique has been considered as a promising tool to monitor the earth surface on global scale continuously. Many researches have been conducted to identify land cover types and estimate the ground heat budget by utilizing remote sensing data, but many problems have not been well addressed: identifying certain vegetation types is very difficult; the seasonality of vegetation affects the classification result, so that the classification for the remote sensing data from different time of the year may turn out to be different and inaccurate; the estimation of radiation, heat flux and vegetation index from remote sensing data may lead to large bias. So, in order to develop an approach to upscale greenhouse gases flux measurement from cropland by utilizing remote sensing, we shall firstly distinguish cropland from other land cover types, and secondly, develop and evaluate the estimation method for some controlling parameters for greenhouse gases flux from cropland, such as radiation, heat flux, Leaf Air Index (LAI). In this paper, methods for identifying cropland and estimating surface flux related parameters will be introduced and evaluated by the ground measurement. Suggestions for further improvement on estimation accuracy will be provided.

 

Tim Kramer - PALSAR/ALOS vs. Landsat TM: A Classification Comparison

The recent availability of satellite based synthetic aperture radar (SAR) imagery has opened up new avenues of remote sensing analysis. Research within the past several years has shown that SAR imagery has the potential to be extremely useful in areas where continuous cloud cover has posed a problem for traditional visual and NIR sensors such as Landsat and MODIS. The Japanese ALOS PALSAR sensor has been shown to be particularly useful for precise regional land coverage observation as well as biomass estimates in densely forested regions. The primary purpose of this study was to explore the capabilities of new ALOS PALSAR satellite images.

 

Michele Trickey - Evaluation of the use of MODIS Emissive Bands to Detect Low-level Tropospheric Inversions over Antofagasta, Chile

The lower troposphere over the coast of northern Chile, Peru, and southern Ecuador is believed to be characterized by inversions throughout much of the year, causing the area’s aridity. However, scarcity of radiosonde data makes it difficult to detect and characterize these inversions, motivating a search for a way to do so through remote sensing. I therefore test for Antofagasta, Chile (23.43° S, 70.44° W) a technique that has shown some success in describing inversion depth and strength over Antarctica. This method is based on the difference between the brightness temperatures of the 7.35 mm water vapor and 11 mm thermal infrared bands (bands 28 and 31) as sensed by the MODIS Aqua platform, and it had relatively high success in predicting the strength and depth of strong inversions over Antarctica. I found that, while the 7.35 mm band temperatures did not accurately represent the brightness temperatures of the atmosphere at the top of the Chilean inversions, the brightness temperature differences between bands 28 and 31 did reflect the seasonal cycle in inversion strength and duplicate some major events in changes of inversion strength. These results, combined with clear paths available to better their accuracy, suggest that it should be possible to develop consistent remotely-sensed descriptions of strong inversions along the South American coast.

 

Angel Hsu - Blue-skies Beijing? Deriving Aerosol Information from MODIS to Assess Air Quality in Beijing

Beijing’s air quality came on center stage when China received the 2008 Olympic bid. Realizing that they had to implement drastic measures to improve the city’s air quality, the Chinese government instituted major pollution control measures in Beijing. These included shutting down factories in neighboring Hebei and Shandong provinces, decreasing the number of vehicles on the road by one-half, halting all construction projects, and increasing public transportation in the months leading up the Olympic Games. While the Chinese government touted the marked improvement these measures and other efforts had in cleaning up Beijing’s air, there were questions as to the credibility of this claim. One study by Andrews claimed that the government overestimated the number of “blue-sky days” based on the Air Pollution Index1 (API) by 22 percent in 2007 and 15 percent in 2008 through statistical manipulation, which suggested that air quality in Beijing was not necessarily improving.

Satellite monitoring and remote sensing could potentially shed light on these discrepancies in Chinese air quality information. Given this context, this study will employ remote sensing techniques to gain a better understanding of how satellite-derived information can be used to monitor air quality in an urban setting like Beijing.

 

Sarah Dewey - Cloud Lines Off of Cape Hatteras

This project examines cloud formation over the Gulf Stream at the point at which its path diverges from the United States’ Atlantic Coast. This change in direction of flow lies near Cape Hatteras, so data is obtained from North Carolinian coastal buoys and from MODIS Terra images of the Cape Hatteras area. On select days in the warmer parts of the year, cloud cover over Cape Hatteras is minimal save for a uniform cloud line along the path of the Gulf Stream. Sea surface temperatures (SSTs) for this area of the Gulf Stream were derived from MODIS Level 1B data and from MODIS Level 2 SST products. Wind direction data was taken from buoys off of the coast of Cape Hatteras, and from radiosonde data in the area. These data together describe the environmental conditions necessary to form a cloud line over the Gulf Stream; in general this investigation is an exercise in combining in-situ with remotely-sensed data to assess these conditions.

 

Meagan Fitzpatrick - Characterize Aedes Mosquito Habitat

Dengue fever virus infects over 50 million people each year. The virus is transmitted through the bite of Aedes aegypti, an urban mosquito. It may be possible to use remotely sensed data to understand the environmental factors which contribute to the distribution of the mosquito. However, the validity of our associations may in turn depend on the choice of satellite used to procure the information. Here, I examine the ways in which images of the same cities from different satellites (Landsat and ASTER) lead to different approximations of vegetative ground cover and built ground cover. I also compare two indices of vegetative ground cover to see if both are sensitive in the same way to changing the satellite. My comparisons suggest that the choice of satellite does matter, but without field data I was unable to definitely determine which gives more representative information.

 

Sarah Guagliardo - Characterization of Mangroves in Cajigal Municipality, Sucre State, Venezuela

Previous studies have identified areas of high and sustained malaria transmission, termed “hot spots,” which were associated with proximity to and quantity of nearby Anopheles aquasalis breeding sites. Other environmental variables associated with high malaria transmission in Sucre State include elevation and slope. It is evident from prior research that the most suitable breeding habitat for An. aquasalis mosquitoes in the pre-adult stage is the Avicenia germinans mangrove species. Preliminary geographic analysis of the study area has suggested that mangroves are farther away from malaria-endemic villages than would be expected. In some cases, mangroves are more than 3 km away from the town centers, which is unusual as it is suspected that mosquitoes cannot fly a distance greater than 1 km.

Mosquito breeding sites identified by ground surveys provide an incomplete picture of the true availability of breeding areas, since it is unfeasible to identify all sites. Remote sensing allows investigators to survey the prevalence of mangroves in what would otherwise be very difficult environment to work in. Accordingly, the purpose of this study is to determine the location of A. germinans and other land cover variables relative to An. aquasalis breeding sites and malaria hot spots using remote sensing technology. Such information is necessary in order efficiently direct malaria control efforts to geographic areas of greatest need.

The objectives of this project are two-fold:
1.) Classify land cover in the areas around towns with high malaria burden using both Landsat and ASTER images
2.) Compare area summary statistics for Landsat and ASTER images

 

Vasso Pappa - Comparison of Landcover Change Detection Techniques

Many techniques have been used in order to detect change based on remote sensing data. In post classification change technique, the images are classified and then compared with each other in order to detect differences among the various categories. Macleod and Congalton used this technique in order to monitor changes in the Zostera marina eelgrass population. Comparing various indices has been also used in order to detect change from remote sensing data. The Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Water index (NDWI) have been used by Owor M. et al. in order to detect seasonal change in the wetlands of Lake George in Uganda. Another technique commonly used in change detection is the Tasseled Cap (TC) transformation. This transformation consists of three different indices- brightness, greenness and wetness. Healy S.P. et al. used the TC transformation in order to detect forest disturbance in St. Petersburg, Russia.

In all of the aforementioned studies the study area was undergoing profound or anticipated changes and the techniques used were aiming to quantify this change. However, the area I will be focusing on is a region that does change neither rapidly nor extensively. Nevertheless, studies on tick populations indicate that that Ixodes Scapularis, the vector for Lyme disease is expanding towards this region. The aim of this study is to use change detection techniques in order to examine what kind of land cover changes have taken place in the study area and to assess which of these techniques is more appropriate to detect subtle changes of land cover.

 

Molly Rosenberg - Landscape Characterization of Risk for Lyme Disease on Block Island: New techniques for creating risk maps in small island settings

Lyme disease is the most common vector-borne disease in the United States and is continuing to spread. The bacterium that causes the disease Borrelia burgdorferi is transmitted via the tick species Ixodes scapularis. The principal host of I. scapularis in its larval and nymphal stage is the white-footed mouse while the white-tailed deer is the principal host in the adult stage. Both of these hosts prefer similar land-cover patterns with a mixture of forest and open areas and these habitats often coincide with suburban residential areas. Remotely sensed data have been used in many studies to create risk maps for vector-borne diseases in general and Lyme disease in particular based on environmental factors.

Block Island is a small island with an area of about 9.7 square miles off the coast of Long Island Sound. Lyme disease is endemic to the island and incidence rates among the population of about 1,010 are high and there are about 40 to 45 new cases every year. A Lyme disease risk map for Block Island will be created based on the landscape characteristics of wetness, greenness and elevation.

 

Brian Seavey - The Potential Use of Landsat TM Imagery to Create an “Early Warning” Praziquantel Supply System to Improve Schistosomiasis Control

Amidst the current global public health outcry against diseases such as HIV/AIDS, malaria and tuberculosis, there exists a myriad of other, neglected tropical diseases that devastate the lives of billions of people worldwide. One such disease, schistosomiasis (sometimes called bilharzia), is of particular importance, as it is responsible for chronic urinary and intestinal infections that can also result in long-term growth and cognitive effects. Schistosomiasis the disease occurs from infection with a Schistosome parasite, which is spread via certain species of freshwater snails. The species of snail varies geographically; however, this project will focus on the Biomphalaria pfeifferi of West Africa. Since snails are the sole vectors of schistosomiasis, understanding their population dynamics is important to predicting risk of infection. If disease risk could be estimated earlier than the presentation of clinical symptoms, the local health facilities in Senegal could prepare effectively to provide care by targeting the population with Praziquantel treatment. In the past 30 years, remote sensing techniques have become an important tool in the epidemiology of many infectious diseases.

To date, there remain many opportunities to explore the use of remote sensing techniques in West Africa in Schistosomiasis control. This project examines snail density in a region of the Senegal River Basin, near the city of Richard Toll, Senegal. Richard Toll (16’38” N; 15’78” W; altitude < 10m) is a small town of around 50,000 – 100,000 people in northern Senegal that has become a regional hub as a result of the large sugar cane irrigation of the Senegalese Sugar Cane Company (CSS) that was started in the early 1980s. Situated along the Senegal River, Richard Toll was greatly affected by the installation of two dams in the late 1980s, the Diama dam near the mouth of the river and the Manantali dam in western regions of Mali. The introduction of these two dams has redefined the nature of the Senegal River, raising water levels and decreasing the salinity. As a result, shortly after the dams became operational, there were Schistosomiasis outbreaks reported in areas where the diseases had previously been nonexistent. Further studies of Schistosomiasis in Richard Toll have shown that the B. pfeifferi snails responsible for spreading the disease are highly seasonal, with some suggestions of certain habitat preferences.

 

Peter Christensen - Examining Agriculture-Urban Land Conversion in the Rio Bravo Watershed

The Rio Bravo basin is experiencing rapid rates of urban growth. This study examines the extent of agricultural-urban land conversion within one rapidly growing irrigation district between 1992-2008. The study investigates the degree to which analysis of land cover change can capture land tenure change as a result of hydrological disturbance. The study compares conventional post-classification change detection techniques to multidate change-vector classification in urban analysis. Results indicate that post-classification techniques may perform more reliably in this study site, while errors could be minimized by identifying regions of change that are confirmed by both methods independently.

 

Norio Takaki - Characterizing Urban Expansion Around Brasilia

Remote sensing technology can be a powerful tool for understanding processes that influence local and regional climate. This study assesses the loss and gain of vegetative cover in an urbanizing landscape and the impact such changes have on surface temperatures. Special attention is dedicated to comparing patterns of change in vegetative cover and surface temperature amongst three different urban areas. Specifically, the primary goals of this investigation are:
1- measure land cover changes that took place in the span of twenty years in an area of 2,229 square kilometers surrounding the Brazilian capital, Brasilia, with a focus on urbanization;
2- calculate mean surface temperatures for different land cover types and compare temperatures amongst different urban areas with varying levels of vegetative cover.

A normalized difference vegetation index (NDVI) was used to track changes in vegetation cover in three primary urban landscapes characterized by different socio-economic make up. The results indicate increases in vegetated cover in higher income neighborhoods from 1986 to 2006, and no significant improvements in vegetation cover for the lower income areas. It was also found out that higher-income urban areas are slightly cooler than their lower-income counterparts (in the year 2006), which is consistent with the corresponding increase in vegetation in the former. The highest mean temperature of all study regions was found in bare soils while the lowest was associated with water surfaces.

 

Tianming Chen - Influences of Urbanization on Local Climate in Beijing

Like most cities in China, Beijing is undergoing fast urbanization and at the same time, its local climate is changing. The most important anthropogenic influences on climate are the emission of greenhouse gases and changes in land use, especially the increase of built-up area and decrease of agriculture and green area. I am trying to explore how the land-use change caused by urbanization, influences local climate in Beijing. Remote sensing could be a good tool to facilitate my exploration, because the satellite images could tell us the morphological changes of a city and temperature information as well.

 

Dea Doklestic - Examining the 2009 Drought Effects in Buenos Aires, Argentina

Climate and weather affect the human population as never before, given its recent increase in both number and density. It is therefore important to understand the effects of weather patterns on the environment and ways in which they can be studied. The phenomenon of recent droughts in Argentina makes a suitable case study. This year, Argentina has been hit by a severe drought, repeatedly labeled as catastrophic. The drought, caused by the meteorological phenomenon “La Niña,” has affected Argentina’s core agricultural producing regions between Rio Parana in the north and Rio Colorado in the south. A better understanding of this occurrence is of both scientific and economic importance.

The effects of the drought are the most easily observable as decrease of vegetation cover, therefore I will be looking into NDVI time series calculated over the region of interest. Another possible consequence of drought is the increase of albedo as the vegetation cover diminishes (observable in areas with bright soils). This phenomenon will also be tested for. All information (other than in-situ measurements of temperature and precipitation) are the result of satellite images obtained through MODIS (MODerate Image Spectrometer) sensor on board the Terra satellite.

 

Aaron Judah - Measuring Changes in Feedback Mechanisms Responsible for Exacerbating Heatwaves and Droughts from Space Observations

The purpose of this project is to use space based measurements to measure changes in feedback mechanisms that are affected during droughts and in very extreme cases can be driving factors in determining the severity and length of a drought. I will accomplish this through the use of MODIS measurements of Normalized Difference Vegetative Index (NDVI), Enhance Vegetative index (EVI), temperature, and surface albedo (near infrared (NIR) and visible). Measuring differences in NDVI will give us a measure of the overall health of the vegetation in our area of interest. The health of the vegetation also drives feedback mechanisms such as albedo, which in turn can change the surface temperature (both factors I am also measuring). By taking measurements of EVI I hope to compare and contrast it with NDVI measurements to see if its enhanced features are useful for these types of studies. With these measurements, I will examine their time series from the spring to summer months, for different years – years with droughts and without droughts.

I also will take coarse measurements of river water level, and precipitation to see if large areas affected by droughts can show trends when coarse measurements are used in comparisons, when contrasting times affected by droughts to times not affected by droughts. The thought being that, coarse measurements during drought conditions, might have enough resolution to act as a measure of the hydrological state of the area. The area I will study is in the Southeastern United States, primarily in the states of Alabama and Georgia.

 

Tamara Machac - Surface Heat Budgets and Land-cover in Northern Wyoming

Land-use and land-cover are major factors in the heat budgets on the regional and local climate scale. In semi-arid and arid regions in particular, changes in water availability dramatically impact the land-cover, which can affect the albedo, surface roughness, the exchange of water, energy and carbon dioxide with the atmosphere, and therefore the surface temperature of the region. This project investigates the accuracy of various methods land-cover classification and albedo for the purpose of calculating accurate surface heat budgets that can vary ground cover.

 

Nicole Thom - The Recent Flood of the Red River in North Dakota

The recent flood along the Red River between North Dakota and Minnesota was the worst flood on record for the area. This paper looks at the extent of that flood and takes into account what type of land was affected. Using MODIS 500m data, the change in overland water coverage was measured. This was done by comparing the water classifications from a pre-flood date (February 2009) to a peak flood date (March 2009). Information about land cover types from July of the previous year was then compared with a cutout of the water class in the March image to determine what types of land were covered by floodwater.