Strength of Forcing Over SM 50 -50 -150 -100 -50 050 100 150 Longitude (°) Temperature GPP Strength of Forcing Over Air Temperature GPP 50 -50 -150 -100-50 050 100 150 Longitude (°) Strength of Forcing Over GPP Temperature 50 -50 Air Temperature 1 0.8 0.6 0.4 0.2 0.2 0.4 0.6 False-Positive-Rate (d) FIGURE 9. (continued ) (c) Convergent cross mapping. An example of the unbiased CCM application used in [88] to derive causal relationships among the variables accounting for photosynthesis [the gross primary productivity (GPP) from FLUXCOM initiative (www .fluxcom.org)], -150 -100 -50 050 100 150 Longitude (°) (c) SM SM SM C (0.7667) Cs (0.7953) 0.8 1 temperature (Tair from ERA Interim), and SM (from the European Space Agency's climate change initiative [(v 2.0)]. Data cubes at 0.5° and eight-day spatial and temporal resolutions, respectively, spanning 2001-2002 were used. Reasonable spatial-causal patterns are observed for SM and Tair on GPP; GPP drives Tair mostly in cold ecosystems (probably due to changes in land surface albedo, such as snow/ice to vegetation changes); SM is mostly controlled by Tair, which partially drives evaporation in water-limited regions; and GPP dominates SM. (d) The additive noise model. Structural equation models in the form of an additive noise model using the kernels in [21] for hypothesis testing. Assessing cause-and-effect relationships is also possible when time is not involved. Here we rely on a look-up table generated by radiative transfer model (RTM), which gives the right direction of causation: state vectors (parameters) cause radiances. The algorithms accurately detect this from pairs of data and can be used for retrieval model-data intercomparison and RTM assessment. 100 IEEE GEOSCIENCE AND REMOTE SENSING MAGAZINE JUNE 2021 Latitude (°) True Positive Rate Latitude (°) Photosynthesis Latitude (°) " "http://www.fluxcom.org http://www.fluxcom.org