Research Component I
Remote sensing and GIS are essential tools for a physical geographer. However, both are quite complex and contain thousands of tools that allow a researcher to analyze various aspects of a landscape, anthropogenic impacts, and the capacity to integrate social and physical science data. Prior to discussing these tools in depth however, it is important to have an understanding of what each is and secondly, why these tools are useful in different applied types of research.
Remote sensing is the acquisition of data from something without actually being in direct contact with it. The most obvious example is a satellite orbiting in space, with a viewing lens pointed towards the earth. Through this lens is a sophisticated type of camera that acquires images of the earth's land, water bodies, atmosphere, vegetation, soils, et cetera. Aerial photography is another example of remote sensing. We can remotely sense not only in the visible part of the electromagnetic spectrum (EMS) (what humans see every day), but also with special equipment we can acquire imagery from space or from airplanes that records information about longer and shorter wavelengths of light. When we capture an image (either digital or analog), what we are actually doing is recording how light interacts with matter. For example, a shirt that appears black is actually absorbing all visible light, and reflecting nothing in the visible range of the EMS. A shirt that appears white is reflecting all light away from the fabric of the shirt. A shirt that is blue is absorbing all light except blue, which is being reflected back to the viewers eyes. Shown below is the entire EMS.
(image is taken from http://rst.gsfc.nasa.gov/Front/overview.html, units are in micrometers, 1x10^-6 SI meters, unless otherwise indicated)
The energy from the sun is measured in wavelengths, other types of energy can also be measured this way. For clarification, here I will describe several of these wavelengths in the chart above. To the far left of the EMS are the gamma wavelengths. Gamma wavelengths are the tiniest wavelengths on this chart, and they have the most energy. These rays get absorbed in our atmosphere and also can help cancer patients, due to the fact that gamma rays kill cancerous (and healthy) cells. For information about gamma rays, please click here. We are all familiar with the X-Ray, these rays can pass easily through living tissue and create an image of our bones, and imperfections (such as a break or fracture). X-rays help remote sensing researchers because they allow us to detect the presence of black holes, which emit this type of energy. For more information describing X-rays, please click here. The next shorter wavelength is ultraviolet (UV). UV rays are what causes sunburns, and on some days more UV light energy gets through our atmosphere than other days. UV light tells us a lot of information about other stars and galaxies. For more information on UV energy, please click here. The infrared portion of the EMS can tell remote sensing researchers a LOT about vegetation (hence my interest!). In combination with the visible portion of the EMS, remote sensing scientists can extract information about plant health, soil moisture (or dryness), invasion of exotic plant species, and soil characteristics, for example. Remember that the part of the EMS that we can see without the aid of instrumentation is seen above. The colors, from shortest to longest wavelength are violet, indigo, blue, green, yellow, orange and red. We see these colors sometimes after a rainstorm in the form of a rainbow.
As we are moving from the left to right on the EMS, wavelengths get longer, the wavelength energy gets lower and the wavelength has the potential to travel further. Remember remote sensing scientists analyze how this energy (wavelengths) interacts with matter, or, in my research, with different types of elements covering the land (like urban areas, agriculture, bare rock, beaches, and many different kinds of vegetation). When wavelength energy interacts with matter, we can record this interaction in several ways. The first is a digital image. Within this image are pixels (picture elements). Pixels are tiny squares within the image that hold a grey-scale value or hue of color. You can see pixels when you zoom in far enough such as the satellite image below:
In remote sensing we try to categorize or, classify each pixel so that it is representative of some element of the earth's land or seascape. We do this by identifying pixels that we know represent water, agriculture, a certain soil type, or a unique rock type, for example. Once we have identified these (representative) pixels and their locations, we train the computer software to look for other pixels throughout the entire image that also fall within some standard deviation of those representative pixel values. For each unique land or seascape, within each pixel or set of pixels that represent that type of land cover, there is something called a spectral signature. Just like how human beings each have a unique signature, so do plants, different water conditions, vegetation types, urban dwellings, et cetera. Once we know what the signature is for example, red mangroves, we can classify an entire satellite image based on a few training pixels in one part of the image. The signature is a function of the percent reflectance of a given wavelength. An example of these types of signatures can be seen below:
(Image is taken from http://rst.gsfc.nasa.gov/Intro/Part2_5.html)
Since the primary satellite image type used in this research is Landsat, this will be the primary focus for the rest of this site in terms of remote sensing. The Landsat sensor gathers reflectance data in certain spectral bands. A spectral band is a range of wavelengths. Landsat has 7 bands. This means that it has 7 distinct bandwidths defined by a range of wavelengths. Band 1 is from 0.45-0.52 µm (blue), band 2 is from 0.52-0.60 µm (green), band 3 is from 0.63-0.69 µm (red), band 4 is from 0.76-0.90 µm (near infrared), band 5 is from 1.55-1.75 µm (mid infrared), band 6 is from 10.4-12.5 µm (thermal infrared), and band 7 is from 2.08-2.35 µm (mid infrared). Remember that each numerical range represents a range of wavelengths. Bands 1, 2, and 3 are in the visible portion of the EMS. The most useful part of the EMS for land cover change analysis is in the visible (band 3) and the infrared (bands 4, 5, and 7). Band 3 is excellent for identifying the red chlorophyll absorption of healthy green vegetation and also for distinguishing among various types of vegetation. Band 4 is useful for identifying difference crops and emphasizes land/water and soil/crop contrasts. Band 5 shows moisture content in plants. Band 7 is useful for delineating between rock formations, which helps researchers locate certain vegetation only found associated with certain rock or soil types.
It is possible to combine several bands at once to retrieve all useful spectral information pertaining to 1 or more land cover types. Different band combinations contribute to a researcher's ability to find the defining characteristics of many types of land cover or more specifically, vegetation species. Not only are the spectral signatures of use, but with some crops, such as citrus plantations for example, there are certain landscape patterns that can be enhanced through the use of band combinations. Pattern identification is also very important when trying to ascertain vegetation type, health, or anthropogenic use.
Remote sensing imagery can be combined with other types of data to create more sophisticated modeling, forecasting, and more accurate estimation of land use/ land cover change over time. An ideal vehicle to do such an analysis is with GIS software. A Geographic Information System is a software that allows for the creation, acquistion, management, analysis of multiple datasets of both vector and raster format. A GIS allows for many data layers to be added simultaneously, mapped, and analyzed. The user asks a question of the data, and the answer is a map.
Both remote sensing and GIS analysis were and are used in Dr. Gebelein's Cuba research due to the many different data sources and format types. A GIS easily allows for this type of integration to take place. Dr. Gebelein's current work and how both remote sensing and GIS are applied to her Cuba research is outlined, discussed and described here.
