CHALLENGES IN USING BRAZILIAN SOIL DATABASES FOR REGIONAL MODELING OF SOIL CARBON AND GREENHOUSE GAS EMISSIONS

introduction

Soil organic matter modeling integrates soil, plant, and climate information to assess the impacts of climate change and mitigation strategies, as well as guide public policies. However, the application of modeling on a regional scale requires key soil data (e.g. C, N, particle size distribution) and vegetation (natural/pre-conversion and cropping systems, when applicable). More often than not, these data are generally lacking or are at least sparse and unevenly distributed in Brazilian soil databases. 

objectives

We evaluated challenges for regional-scale modeling of C and GHG based on the DayCent model, in particular data availability in Brazilian soil datasets.

results

Soil data sources in Brazil (mainly EMBRAPA and SoilData) usually reflect the limitations of soil surveys: coverage of the national territory is limited to small scales (1:250,000 at most), with uneven geographical distribution of the sampling site and large sampling "gaps". For the purpose of modeling, the metadata is often incomplete, with generic information about the vegetation (or cropping systems) on site. The standardization of analytical methods and communication of data quality (uncertainties) is often incomplete. Databases are usually structured according to pedological profiles described in the field (by soil horizons) or targeted surveys using specific layers (0-10 or 0-20 cm depths), which are not directly imported into the DayCent model using hard-set layer structures. These incompatibilities can be resolved by harmonization techniques, but those require additional computational effort and some level of programming skills.

implicatio

The Pronasolos (National Program for Soil Survey) is a pivotal opportunity to improve spatial scales and obtain up-to date soil data wich is crucial to biogeochemical cycle modeling. This requires coordination between project planners, field teams, data scientists and modelers; key variable monitoring projects (SOC, for example) should be implemented on a national scale, considering site-specific aspects (at biome level?). Brazilian science still depends heavily on information from other countries to understand its territory; new technologies that should give a new boost to modeling are the interaction with AI and new data sources, such as the new Biomass satellite;

There is a pressing demand for advanced soil C and GHG modeling in Brazil, which could provide a crucial contribution to the National Inventory of Green House Gases. However, in view of our findings, there is an essential need for expanding and refining soil and vegetation databases --- an instigating challenge for Brazilian Academia, and public and/or private initiatives as well.