Supporting Online Material for

Journal of Geophysical Research: Biogeosciences

Supporting Information for

Attribution of land-use/land-cover change induced surface temperature anomaly: how accurate is the first-order Taylor series expansion?

Chi Chen1*, Liang Wang1, Ranga B. Myneni1, Dan Li1*

1Department of Earth and Environment, Boston University, Boston, MA, USA

Contents of this file

Text S1

Figures S1 to S2

Introduction

This supporting information contains two supplementary figures. Text S1 describes the CESM outputs. Figures S1 and S2 present the biases of the estimated changes in LST using CESM outputs as the forcing for TRM.

Text S1.

To verify the spatial patterns in Figures 4 and 5, we use results from simulations with the Community Earth System Model (CESM) Version 2 to force the FOTSE and SOTSE LST and attribution models. The simulations are offline simulations with the Community Land Model (CLM) Version 5 (Lawrence et al., 2018), which is the land component of CESM2, and driven by prescribed atmospheric forcing (called GSWP3, Global Soil Wetness Project), as well as the climatology of leaf area index (LAI) derived from MODIS satellite. The LAI has monthly variations but no year-to-year variability. This simulation starts with initial conditions that have already been spun-up. We also discard the first ten years of simulation. The reference land cover is 2000, and the atmospheric CO2 level is fixed to 367 ppmv. We calculate the annual average of each climate variable and then take their mean from 2000 to 2017 as the final forcing for the TRM framework. The spatial resolution is 0.47° × 0.63°.

References

Lawrence, D., Fisher, R., Koven, C., Oleson, K., Swenson, S., Vertenstein, M., et al. (2018). Technical Description of version 5.0 of the Community Land Model (CLM), 1–350.

Figure S1. Similar to Figure 4 but using CESM outputs as the forcing for TRM. Biases of the estimated changes in LST in the small perturbation case, where =0.01, =10 s m-1, =10 s m-1.

Figure S2. Similar to Figure 5 but using CESM outputs as the forcing for TRM. Biases of the estimated changes in LST in the large perturbation case, where =0.05, =50 s m-1, =50 s m-1.

1