TOWARDS QUANTIFYING SYSTEM RESILIENCE FOR CLIMATE-SMART AGRICULTURE AND FORESTRY

The climate-smart agriculture (CSA) initiative pursues the triple wins: (1) sustainably increasing productivity and incomes; (2) adapting and building resilience to climate change; and (3) mitigating greenhouse gases emissions. CSA provides the practical aims that facilitate the development of holistic indicators for quantitative evaluation and monitoring, on which decision-making support system will be based. In this endeavor, quantitative assessment of resilience is one of the main challenges not only in agriculture but also in forestry. We estimated normalized spectral entropy as a proxy measure of subsystem resilience using decade-long tower-based fluxes of energy and matter over forest and agricultural ecosystems in Korea. These ecosystems were examined to find out whether the current state of ecosystem management is a proper configuration toward climate-smart agriculture and forestry. The study sites are two adjacent forests (i.e., a fast-growing coniferous forest plantation and an aged deciduous forest) in Pocheon, and a heterogeneous annual cropland in Haenam, Korea. Eddy covariance technique was used to measure fluxes of carbon, water and energy along with other micrometeorological, plant and soil variables from 2006 to 2015. The 30-minute data were processed and gap-filled by the KoFlux protocol. Spectral entropy (H_s) was calculated from power spectral density and then normalized (H_sn) to have a range from 0 to 1. If H_sn is close to 0, it implies that the time series is composed of periodic cycles and varies orderly. In this context, 1- H_sn may be used as a measure for evaluating the adaptive capacity of the subsystem, i.e., a key mechanism for resilience. We then selected a suite of variables associated with biophysical and biogeochemical processes in these ecosystems, including radiation components, precipitation, evapotranspiration, gross primary productivity, ecosystem respiration, net ecosystem exchange of CO₂, air temperature, canopy temperature and soil temperature. The spectral entropy values of these variables are related and compared for different periods and are interpreted to link with the self-organizing processes in these contrasting ecosystems. Our results demonstrate that these ecosystems manifest self-organizing processes which may be used to gauge the ecosystem resilience.