Remote sensing is a technology used to gather data about an object or event from a distance. This involves both airborne and satellite imagery, plus the image processing necessary to develop products of value for land management, monitoring, and analysis. Remote sensing data consists of everything from traditional aerial photography to high-resolution digital satellite imagery, and can include active systems such as radar and lidar. Remotely sensed data has many applications and uses. Among them are mapping wildland fires, providing three-dimensional perspectives of landscapes, delineating wildlife habitats, and monitoring vegetative conditions.
Selecting the Map Level Appropriate for Your Project
Maps are developed and used at multiple resolutions and are best represented by a hierarchical series of mapping products. These products are described as map levels. Four hierarchical map levels that represent a gradient of thematic and spatial resolutions are regularly used to bracket the scale and use of a map product. Table 1.1 of the Existing Vegetation Classification and Mapping Technical Guide Version 2.0 illustrate the business requirements and applications; Table 1.2 presents characteristics of these map levels. The included link is a current version of the manual, and references to Tables might have changed. The four map levels are as follows:
- National. National, the coarsest level in the map hierarchy, is intended to store and depict data on a nationwide or global scale. Map products at this level will typically have broad classes and coarse spatial representation. Products at this level may be developed programmatically or aggregated from existing lower-level products where feasible.
- Broad. Broad-level products are intended to support state, multistate, or regional information needs. Products at this level may be developed programmatically or aggregated from existing products.
- Mid. Mid-level products are intended to support forest and multiforest information needs, including forest planning, forest/region resource assessment and monitoring, and fire/fuels modeling. Products at this level provide a synoptic and consistent view of existing vegetation across all ownerships within the map extent. They are usually developed programmatically from remotely sensed data and field data.
- Base. Base-level products support local forest and district information needs, and represent the highest thematic detail and spatial accuracy. Most project planning and implementation require products at this level. Due to the cost of development, base-level information is the least likely of all levels to be spatially extensive; however, it offers the most flexibility for upward integration within the map hierarchy. Products at this level are typically developed from large-scale, remotely sensed data and field data.
Because riparian areas tend to be relatively narrow with linear features, most riparian projects will use the mid- or base-level maps.
Selecting Imagery Appropriate for Your Project
There is a tremendous variety of remotely sensed imagery–with differing scales, spatial resolution, spectral sensitivity, revisit time, and other variables. Having a multitude of choices is good, but it requires that we understand the capabilities and limits of available imagery in order to use it appropriately.
Types of Imagery
Following are examples of sensors with varying properties categorized into spatial resolution classes (from high to coarse).
High Resolution Imagery (spatial resolution of 10 m or less). Examples of imagery include resource photography (film or digital), DOQs, lidar, and some satellite images.
Medium Resolution Imagery (spatial resolution between 10 and 30 m). Examples include SPOT, ASTER, Landsat TM and ETM+.
Medium-Coarse Resolution Imagery (spatial resolution between 30 and 250 m). Examples include: Landsat MSS, AWiFS, and MODIS (250 m).
Coarse Resolution Imagery (spatial resolution greater than 250 m). Examples include MODIS (> 250 m), and AVHRR.
Matching Information Needs with Imagery
To successfully use remote sensing, information requirements must be matched with imagery that can deliver the information effectively. The first step is to understand the information requirements. For example: How large is the geographic area? What level of detail and spatial accuracy is needed? In general there is an inverse relationship between area coverage and spatial resolution. In this respect, riparian areas present a special challenge. Riparian areas tend to be relatively narrow but long, extending over great distances. This means that a riparian investigation may require imagery that covers a large geographic area with a sufficiently high resolution. This can result in high costs for imagery acquisition.
In addition, one should know–or at least consider–how often the information is needed. Is it only once, once per year, seasonally, or daily? Satellite sensors have varying revisit times, which may be an important factor. Of course, draped over all other considerations, is cost–both financial and consequential. What is the cost of the data? How much will be spent on the processing? And what is the cost of not having the information?
In a perfect world we would have access to imagery that has high spatial resolution, covers a large area, has all of the spectral data necessary to discriminate our features of interest, and is free and easy to use. Unfortunately, most of the time, these image attributes conflict. Thus, in the real world we must compromise and make an informed choice from what is available.