The PnET Models: Carbon, Water and Nitrogen Dynamics

• Investigators: John Aber, Mary Martin, Scott Ollinger
• Contact: John Aber
• Start date:
• End date:
• Location:
• Latitude:
• Longitude:
• Elevation:
• Taxa:
• Keywords: air pollution, carbon, climate change, GIS, models, nitrogen, ozone, water
• Abstract:

  The PnET models are a set of nested algorithms describing the interactive cycles of carbon, water, nitrogen and additional elements in forest ecosystems. It is what is a “lumped-parameter model" in that it has been designed to capture the important features of forest ecosystem function while requiring the as few input parameters as possible. It is also a non-calibrated model, in that input parameters describing physical, chemical or biological processes (e.g. foliar N concentration, solar radiatioin) are derived directly from field measurements, and are not modified to increase agreement between model predictions and ecosystem-level measurements (e.g. photosynthesis, streamflow). PnET is an open-source model. Native code and compiled versions, along with supplemental information and papers describing the structure and applications of the model are available at http://www.pnet.sr.unh.edu.

  The core algorithms (PnET-Day) predict canopy level carbon balances and potential evapotranspiration. At the heart of this system is a relationship between the nitrogen concentration in leaves and the maximum rate at which photosynthesis proceeds. This rate can be altered by atmospheric concentrations of CO2 or ozone (Ollinger et al. 1997). Species are represented here and throughout PnET by a set of physiological parameters describing foliar N content, specific leaf weight, foliar retention time etc. By using different foliar nitrogen values representative of different species, students can test the effects of species composition on carbon fixation. This version of the model has been tested against gross photosynthesis measured by eddy covariance at the Harvard Forest (Aber et al. 1996) and is now being modified to use as part of the Ameriflux network.

  PnET-II adds a complete water and carbon budget to the photosynthesis routines in PnET-Day. Water use efficiency (WUE) is a function of vapor pressure deficit, and transpiration is WUE times gross photosynthesis. Foliar, wood and root allocation and respiration are defined by simple, physiologically-based equations. PnET-II has been tested against water yield and NPP data at Hubbard Brook, the Harvard Forest, and against regional data planes for water and carbon balance (Aber et al. 1995, Ollinger et al. 1998).

  PnET-CN adds nitrogen cycling to PnET-II and requires data on nutrient contents in different plant tissues, as well decomposition rates for litter fall and turnover rates for different biomass fractions. This version has been tested against the nitrate leaching loss record from all of the control and experimental watersheds at Hubbard Brook (Aber and Driscoll 1997), and has been used to successfully separate the effects of climatic variation, atmospheric deposition, and physical disturbance in this nitrate signal (Aber et al. 2002). It has also been used to determine the separate and interactive effects of ozone, N deposition, CO2 enrichment and prior land use history on current carbon balances (Ollinger et al. 2002).

  PnET-BGC adds soil chemical processes and the cycling of all major elements to the PnET-CN framework. Weathering rates and soil physical characteristics are added to soil biological processes, and an iterative, thermodynamically-based solution to several soil chemical equilibrium equations is derived. PnET-BGC has been tested against the sulfate and cation stream records for the control watershed at Hubbard Brook (W6, Gbondo-Tugbawa et al. 2001, 2002) and is being applied in scenario testing to evaluate the effects of previous and proposed clean air initiatives on stream water chemistry.

  The Pnet model is a simple and widely-used model of ecosystem processes as affected by changes in climate and air pollution. It is what is termed a “lumped-parameter model" in that it has been designed to capture the important features of forest ecosystem function while requiring the as few input parameters as possible. While versions of the model exist which predict complete cycles of water, carbon and nitrogen (Aber et al. 1995, 1996, Aber and Driscoll 1997), there is also a version which focuses only on photosynthesis; the process by which forest canopies convert sunlight into carbohydrates. In this version, students can test the effects of increased atmospheric CO2 and air pollution (ozone or “smog") on this most fundamental process for forests.

  At the heart of the model is a relationship between the nitrogen concentration in leaves and the rate at which photosynthesis proceeds. By using different foliar nitrogen values representative of different species, students can test the effects of species composition on carbon fixation. Increased CO2 in the atmosphere is also known to increase rates of photosynthesis. By altering this critical global change parameter, students can test the feedback between higher CO2 concentrations and increased photosynthesis. In contract, ozone also has a well-known, but negative effect on photosynthesis. Testing the interactive effects of these two gases on plant processes will allow an introduction to the complexities of global change.

  Photosynthesis requires the use of water and shortages of water can reduce carbon gain. PnET contains a simple water balance routine through which the effect of changes in precipitation, as a part of climate change, can be predicted. Warmer temperatures and a longer period of leaf display in most temperate forests can also affect total carbon gain and water stress. Pnet predicts canopy phenology (the production and loss of foliage) and so addresses this aspect of change as well.

  Taken together, PnET allows students to deal with changes in CO2, ozone, temperature, precipitation and species composition all with a model in which requiring only 17 input parameters (of which only 6 vary between sites or species) plus climate data. The model has been tested extensively against measured water and carbon balances in several forest ecosystems. It is currently included as one of two models in an international project (Ameriflux) in which models are tested against measured data on carbon flux from a series of ecosystems across the Americas.

• Methods:

  Non-calibrated, data-intensive lumped-parameter modeling approach. Uses regression-based summation of physiologically-based processes linked at time scales of days to months. Models are tested extensively against independent, ecosystem-level responses and then used to predict ecosystem responses to future environmental scenarios.

• Related datasets: