To define relative spectral response, I referred to EFFECT OF RELATIVE SPECTRAL RESPONSE ON MULTI-SPECTRAL MEASUREMENTS AND NDVI FROM DIFFERENT REMOTE SENSING SYSTEMS by David James Fleming, 2006.
One factor that is often overlooked is the effect of a sensor’s relative spectral response (RSR), or spectral response function (SRF), on broadband spectral measurements. The RSR describes the quantum efficiency of a sensor at specific wavelengths over the range of a spectral band. Currently, general descriptors, such as bandwidth and average bandpass, are often the only spectral characteristics considered in analysis of sensor spectral measurements. However, cross-sensor wavelength variations in RSR can lead to measurement discrepancies between sensor measurements that make them not directly comparable (Teillet et al., 1997). In order to provide consistent quantitative spectral measurements of vegetation land cover and derived metrics, such as spectral vegetation indices, the effect of a sensor’s SRF must be considered and understood.
Spectral bands are often generalized (Pagnutti et al., 2003) in terms of full width at half maximum bandwidth and central wavelength corresponding to the maximum value of the response function (Liang, 2004) as shown in Figure 11.
In view of this, I was a little bit surprise that, for example the bandwidth of MODIS band 7 (dotted line) as defined here, seems to do not match with the relative spectral response in the way defined above.
Is there something that I did not correctly understand in those concepts?