Evaluating the need for integrated land use and land cover analysis for robust assessment of carbon-related climate adaptation and mitigation strategies

Several climate adaptation and mitigation strategies incorporate Land Use and Land Cover Change (LULCC) to address global carbon balance. However, LULCC is not consistent across the CMIP5 model simulations because only the land use input was harmonized. The associated LULCC uncertainty impedes understanding of global change because LULCC influences regional and global carbon and climate dynamics. For example, the integrated Earth System Model (iESM) overestimates 2004 atmospheric CO2 concentration by 14 ppmv, and we explore the contribution of historical LULCC uncertainty to this bias in relation to the effects of CO2 fertilization, climate change, and nitrogen deposition on terrestrial carbon. Using identical land use input, a chronologically referenced LULCC that accounts for pasture, as opposed to the default year-2000 referenced LULCC, increases this bias to 20 ppmv because more forest needs to be cleared for land use. Assuming maximum forest retention for all land conversion reduces the new bias to 19 ppmv, while minimum forest retention increases the new bias to 24 ppmv. Similar atmospheric carbon differences, however, have different global carbon dynamics. The land carbon difference is larger between the two forest retention cases than between the default and maximum forest cases, with the ocean compensating for this carbon rather than the atmosphere. Nonetheless, the land component gives a 33 Pg land carbon uncertainty range due to maximizing versus minimizing forest area, which is 80% of the estimated 41 PgC gain in land carbon due to CO2 fertilization combined with climate change from 1850-2004 and 150% of the estimated 22 PgC gain due to nitrogen deposition. These results demonstrate that LULCC accuracy and uncertainty are critical for estimating the carbon cycle, and also that LULCC may be an important lever for constraining global carbon estimates. Furthermore, differences in land conversion assumptions generate regional differences of up to 0.8 °C between the two forest retention cases. These temperature differences are the combined effects of changes in general circulation and local land cover. This work highlights the need for more accurate LULCC scenarios with quantified uncertainty in earth system simulations in order to provide robust historical and future projections of carbon and climate.