Figure 6: Weights of the daily data of each sensor according to the temporal distance by the nearest sensing day. The weights are distributed with a gaussian function River. 6.4 River discharge estimation from reflectance indices The estimation of river discharge from reflectance indices is very similar to the rating curve approach applied for the water levels by altimetry. Indeed, here the relationship is based on the evaluation of a non- linear regression relationship between the multi-mission time series and the observed river discharge values.
After processing the multispectral images from the multi-mission satellites following the methodology explained in Section 5.3, the retrieved signals (hereafter denoted as CM signals) are further utilized to derive river discharge time series along the selected river reaches. To formulate the river discharge algorithm, we need to calibrate the CM signals against the contemporary in situ river discharge for any typical river sites. However, along the selected river sites, the in situ observations are often unavailable during the period in which satellite data are available (2006-2005 for Landsat 5, 2021-2022 for Landsat-9, 2016-2022 for the remnants).
Therefore, two different analyses are carried out depending on the availability of the in situ data: i) calibrated approach (when coincident observation of in situ Q and CM signals are available) and ii) uncalibrated approach (when only in situ observation non-contemporary to satellite data are available). 6.4.1 Calibrated Approach With the coincident data availability of CM and Q, River discharge is estimated through the use of Empirical Formulation . Notably, the calibrated procedure is performed following the pre-fixed calibration and validation period.
Best-fit approach In the case of empirical formulation, four potential distributions (linear, quadratic, power, and exponential) are selected as potential laws between Q and CM data as follows: Eq.14a Eq.14b Eq.14c Eq.14d D.3 River Discharge ATBD. Reference: RD_CCI-0006-ATBD - Issue 2.0 – 31/07/2025 Open/Public © 2019 CLS. All rights reserved. Proprietary and Confidential. 27/31