Hostname: page-component-7b7567c8c-7pwfn Total loading time: 0 Render date: 2026-11-16T11:13:26.664Z Has data issue: false hasContentIssue false

Operational sub-pixel snow mapping over the Alps with NOAA AVHRR data

Published online by Cambridge University Press:  14 September 2017

Nando Foppa*
Affiliation:
Remote Sensing Research Group, Department of Geography, University of Bern, Hallerstrasse 12, CH-3012 Bern, Switzerland
Stefan Wunderle
Affiliation:
Remote Sensing Research Group, Department of Geography, University of Bern, Hallerstrasse 12, CH-3012 Bern, Switzerland
David Oesch
Affiliation:
Remote Sensing Research Group, Department of Geography, University of Bern, Hallerstrasse 12, CH-3012 Bern, Switzerland
Florian Kuchen
Affiliation:
Remote Sensing Research Group, Department of Geography, University of Bern, Hallerstrasse 12, CH-3012 Bern, Switzerland
Rights & Permissions [Opens in a new window]

Abstract

This study is part of research activities concentrating on the real-time application of the U.S. National Oceanic and Atmospheric Administration (NOAA) Advanced Very High Resolution Radiometer (AVHRR) sensor for snow-cover analysis of the European Alps. For mapping snow cover in heterogeneous terrain, we implement the widely used linear spectral mixture algorithm to estimate snow cover at sub-pixel scale. Principal component analysis, including the reflective part of AVHRR channel 3, is used to estimate fractions of “snow” and “not snow” within a pixel, using linear mixture modeling. The combination of these features leads to a fast, simple solution for operational and near-real-time processing. The presented algorithm is applied on the European Alps on 17 January 2003 and successfully maps snow at sub-pixel scale. The detailed snow-cover information makes it easy to recognize the complex topography of the Alps, more so than with either a classic binary map or a Moderate Resolution Imaging Spectroradiometer (MODIS) snow product. The sub-pixel algorithm reasonably identifies snow-cover fractions in regions and at altitudes where neither the classic binary map nor the MODIS algorithm detects any snow. Differences concerning the snow distribution are found in forested areas as well as in the lowest-elevation zones. The algorithm substantially improves snow mapping over complex topography for operational and near-realtime applications.

Information

Type
Research Article
Copyright
Copyright © The Author(s) [year] 2004
Figure 0

Fig. 1.Two-dimensional scatter plot using the first two principal components from NOAA-17, 17 January 2003. The extreme points, surrounded by circles, bound the cluster of pixels in the feature space.The corresponding pixels are defined as image end-members following the concept of convex geometry.

Figure 1

Fig. 2.End-member spectra from convex hull (NOAA-17, 17 January 2003).

Figure 2

Fig. 3.Histogram showing the frequency of occurrence of snow-cover fraction values from the NOAA sub-pixel composite image acquired on 17 January 2003.

Figure 3

Fig. 4.Snow-cover fraction daily composite map as a result of the applied sub-pixel algorithm on three NOAA overpasses (N12, 14:42 UTC; N16, 13:19 UTC; N17, 11:00 UTC) from17 January 2003. Most of the values result from the NOAA-16 unmixing output because of the great number of cloud-free pixels over land. NOAA-17 and -12 contribute about 20% to the daily sub-pixel snow map. The image shows the whole European Alps (44–49° N, 5–15° E) covering an area of approximately 600 km by 800 km. The snow-cover fractions have been aggregated into four classes; the 0–15% class represents snow-free areas as discussed in section 3.5. Pixels corresponding to clouds and water bodies are masked out and occur in a pattern signature.

Figure 4

Fig. 5.Snow-cover products and a MinRMS image generated from 17 January 2003: (a) MODIS daily snow product (MOD10A1) at 500 m resolution, co-registered to the AVHRR dataset; (b) binary snow-cover composite from separated NOAA-16 and -17 unsupervised ISODATA classification; (c) subset of the snow-cover fraction daily composite map; (d) MinRMS error map, appropriate to (c). The region of interest is situated in the central Alps and covers an area of approximately 250 × 250 km2. The subset is mostly situated in western Switzerland, characterized by a wide range of elevation zones including different land-cover types. The northwest part of the image is dominated by the Jura mountains (about 1500 m a.s.l), with the lower altitudes of the Swiss Plateau (about 600 m a.s.l.), ending in the foothills of the higher central Alps (up to 4500 m a.s.l.), covering the middle part of the subset. The dotted line in (b) and (c) represents the transect shown in Figure 6. Encircled areas correspond to specific areas of interest which are further explained in section 4.

Figure 5

Fig. 6.Transect crossing the Bernese Alps (A) to Lake Maggiore (B). (a) The altitude along the transect; (b) the snow-cover fraction of the proposed sub-pixel method (solid line) and the MODIS snow cover (dotted line).