Abstract

This study investigates the evolution of the energy, water, and carbon cycle due to land-use change at Calhoun critical zone (CCZ) using field observations and modeling results of eco-hydro-meteorological variables. Land-use change such as deforestation, cultivation, and reforestation alters the energy, water, and carbon cycle by changing land surface temperature, heat fluxes, evapotranspiration, and carbon storage. CCZ is an ideal platform for investigating this issue as Calhoun has experienced a huge land-use change including severe deforestation in the 18th century, intensive cultivation before the Great Depression in the 1930s, and tremendous reforestation in the 20th century.

This study selects current Duke Forest as Calhoun’s pre-agricultural ecosystem before cultivation, current cropland as Calhoun’s agricultural ecosystem during the plantation era, and current young pine forest as Calhoun’s post-agricultural ecosystem since the reforestation movement, and designs a thorough instrumentation to measure land surface energy, water, and carbon fluxes among other eco-hydro-meteorological variables. Three field sites have been constructed at CCZ from August 2016 to May 2017. The three field sites are the above- and below-canopy flux tower sites at the regrow young pine forest, and the cropland site at a hunting field. The observational system measures more than 300 variables from 7 m below ground to 9 m above ground and records more than 500 GB raw data.

The field observations are used to test three models of land surface fluxes. The first model is the maximum entropy production (MEP) model of heat fluxes. The MEP model is tested using half-hourly data of more than seven months at CCZ field sites. The MEP model estimates heat fluxes at all sites accurately with the relative errors no greater than 9%.

The second model is a half-order derivative (HOD) model of gas fluxes that does not use bulk gradient of gas concentration, surface wind speed, or surface roughness. The HOD model is tested using field observations at six field sites with different climate and vegetation types. The results suggest that the HOD model is able to capture the diurnal and seasonal variations of CO₂ fluxes.

The third model is an extreme solution model (ESM) of friction velocity over land surface. The model formulation is based on the extreme solution of the Monin-Obukhov similarity equations relating friction velocity directly to sensible heat flux that may be parameterized using the MEP model. Case studies using half-hourly data validate the model and indicate the possibility for estimating friction velocity and surface wind speed using remote sensing only observations.

The three models are used to fill the data gaps of the eddy-covariance heat and CO₂ fluxes at CCZ field sites. With the MEP model, more than 8000 data gaps of the heat fluxes are filled at the below-canopy flux tower site. With three models together, almost 2000 data gaps of the CO₂ flux are filled at the Duke Forest site. The MEP model is also used to estimate heat fluxes at the cropland site as the eddy-covariance system is not available due to safety concerns.

With the field observations and modeling results of the eco-hydro-meteorological variables at CCZ ecosystems, the evolution of the energy, water, and carbon cycle due to land-use change is quantitatively investigated. The key findings are:

• At long time scale, deforestation increases air temperature (Ta), soil temperature (Ts), sensible heat flux (H), soil heat flux (G), water vapour density (Cv), and evapotranspiration (E), and decreases surface soil moisture (θ). Reforestation decreases Ta, Ts, H, G, Cv, and θ, and increases E. Note that both deforestation and reforestation increase E and decrease θ. The post-agricultural ecosystem has greater Ta, Ts, Rn, H, Cv, E, carbon intake, and water use efficiency (WUE), and smaller G and θ than the pre-agricultural ecosystem. The influence of deforestation on microclimate change is generally greater than that of reforestation.
• At seasonal scale, deforestation increases seasonal variations of Ta, Ts, H, G, Cv, and E, and decreases seasonal variations of θ. Reforestation decreases seasonal variations of Ta, Ts, soil temperature at 30 cm depth (Ts@30cm), H, G, Cv, and E. Land-use change alters Ta, Ts, H, G, and Cv more significantly during the summer than during other seasons. Deforestation alters E more significantly during the summer, and reforestation alters E more significantly during the spring. Land-use change at CCZ alters CO₂ more significantly in winter and CO₂ flux (Fc) more significantly in spring, respectively. The post-agricultural ecosystem has greater seasonal variations of Ta, Ts, Rn, H, G, Cv, E, and CO₂, and smaller seasonal variations of θ and Fc than the pre-agricultural ecosystem.
• At sub-daily scale, deforestation increases diurnal variations of Ta, Ts, H, G, Cv, and E. Reforestation decreases diurnal variations of Ta, Ts, Ts@30cm, and Cv, and increases diurnal variations of H and E. Note that both deforestation and reforestation increase the diurnal variations of H and E. Land-use change alters almost all variables more significantly during the daytime than during the nighttime. However, throughout the year, land-use change alters daytime and nighttime CO₂ equally strong. The post-agricultural ecosystem has greater diurnal variations of Ta, Ts, Rn, H, G, Cv, E, CO2 and Fc than the pre-agricultural ecosystem.
• The differences of the pre- and post-agricultural ecosystems in vegetation types and maturity are the major reason for the differences of the two ecosystems in seasonal cycles of H, CO₂, Fc and WUE.

This study draws three conclusions. Firstly, the flux variables such as H, G, E, and Fc are seven times more significantly influenced by land-use change than the corresponding meteorological state variables such as Ta, Ts, Cv, and CO₂. Although the state variables including Ta, Ts, Cv, and CO₂ at pre- and post-agricultural ecosystems are almost identical, the flux variables including H, E, Fc and WUE at the post-agricultural ecosystem are almost twice of those at the pre-agricultural ecosystem. Secondly, deforestation alters land surface variables six times more than reforestation does. Thirdly, land-use change alters soil conditions such as Ts, G, and θ three times more than the corresponding air conditions such as Ta, H, and Cv. Therefore, microclimate change would be underestimated using the most concerned and commonly measured variables such as air temperature, relative humidity, and CO₂ concentration.

Citation

Tang, Yao (2018): An Observational and Modeling Study of the Energy, Water, and Carbon Cycle at Calhoun Critical Zone. PhD Dissertation, Civil Engineering, Georgia Institute of Technology, Atlanta, Georgia.