Environmental Engineering Case Study - Apollo Mapping

Environmental Engineering

Arcadis is an international engineering, consultancy and design firm with over 15,000 professionals solving water, infrastructure and environmental challenges. In late 2011, Apollo Mapping assisted Arcadis with its need for high resolution satellite imagery in support of an environmental project.

Working on a rural site in New Mexico for a local mining company, this multi-decade project involves a variety of players including government regulatory agencies, resource users and the local community. Arcadis works closely with these groups, balancing their individual needs in support of the larger project goals of environmental restoration, monitoring and remediation of past and current mining activities. Of particular concern is wind-blown dust and mine tailings given the open, rural landscape, dry weather conditions and common wind gusts.

Natural color 80-cm IKONOS imagery collected over Arcadis’ rural site in New Mexico, USA.

Geospatial Data and Spatial Analysis

Spatial analysis utilizing a variety of geospatial datasets and technologies is an important component of Arcadis’ overall decision making process for this environmental project. Apollo Mapping was contracted to provide high resolution satellite imagery from the vendor that could deliver newly tasked data the quickest during a narrow section of the growing season. IKONOS was selected for this project as Geoeye provided us with the shortest tasking feasibility at one month. GeoEye obtained and delivered a cloud-free bundle of georeferenced IKONOS 3.2-meter 4-band multispectral and 80-centimeter panchromatic imagery less than two weeks after order confirmation.

Upon receipt of the raw imagery, Arcadis processed the data with ERDAS IMAGINE to create a variety of layers for further spatial analysis. By fusing the multispectral and panchromatic data in a process called pansharpening, they were able to create natural color imagery with 80-cm resolution. Orthorectification of both the pansharpened and raw IKONOS imagery allowed Arcadis to match the 2011 imagery to historic datasets – a step that is important for assessing changes in ecological communities through time. ERDAS was also used to calculate several indices that assess plant health, including the Normalized Difference Vegetation Index (NDVI) and the Transformed Normalized Difference Vegetation Index (TNDVI).

False color 80-cm IKONOS imagery collected over Arcadis’ rural site in New Mexico; red areas are healthy vegetation.

For the next steps in the analysis, Arcadis employed Esri ArcMap 10.0 for its ability to display and work with multiple spatial datasets simultaneously. Visual inspection of the pansharpened natural color IKONOS imagery was used to map the extent of vegetation communities. These digital maps (i.e. vectors) were compared to similar maps that were extracted from historic datasets in order to assess change through time. The NDVI and TNDVI layers created with ERDAS can reveal areas of unhealthy vegetation that might not be visible with the human eye. Unhealthy vegetation can be an early indicator of environmental damage that warrants additional study and remediation.

The true power of ArcMap for spatial analysis is its ability to combine multiple spatial layers with unbiased statistical and mathematical approaches. These approaches, which include cluster analysis, geographically weighted regression and spatial overlays, can reveal patterns in data that might be missed by visual inspection alone. In their project, Arcadis used a wide variety of spatial layers – e.g. cultural data, current and past vegetative health, soil chemistry – to identify areas that: were healthy and stable; had improving health; and those that were at-risk of further environmental degradation. The results of this spatial analysis have been key to making informed decisions on where to focus remediation efforts and where additional field research is required.

Mini-Glossary of Geospatial Terms

A list of the key geospatial terms used in this case study:

  • Cluster Analysis – a statistical technique in ArcMap that allows users to identify areas that are similar in some sense
  • Geographically Weighted Regression – a statistical technique in ArcMap that allows users to identify trends in observations and processes that may be spatial in nature
  • Hyperspectral Datasets – imagery with multiple bands of spectral data covering narrow frequencies; a traditional multispectral dataset has 4 spectral bands each covering a frequency range of a hundred to several hundred nanometers, versus hyperspectral data which can have 256 spectral bands each covering only a single nanometer of the spectrum
  • NVDI – a simple analytic technique to assess the health of vegetation, it is calculated as:  (Near Infrared – Red) / (Near Infrared + Red)
  • Object-oriented Classification – an advanced remote sensing technique that identifies clusters of similar pixels based on spectral and spatial characteristics as opposed to traditional unsupervised classifications which identify similar pixels based only on their individual spectral characteristics
  • Spatial Overlay – a mathematical technique for spatial analysis in ArcMap whereby users assign weighting values to a set of spatial layers and then sum these values for a comprehensive ranking system
  • TNVDI – an alternate calculation of plant health with fewer negative values:  Square Root [ (Near Infrared – Red) / (Near Infrared + Red) ]