The Data Assimilation Linked Ecosystem Carbon (DALEC) model is a simple box model of the terrestrial carbon cycle.
DALEC v1 is a simple box model of carbon pools connected via fluxes running at a daily time-step. For the evergreen model there are five C pools representing foliage (Cf), woody stems and coarse roots (Cw), and fine roots (Cr) along with fresh leaf and fine root litter (Clit) and soil organic matter and coarse woody debris (Csom). In the deciduous model there is an additional labile pool (Clab) of stored C to support leaf flushing. The following assumptions were made to determine the fluxes between the C pools:
1. All C fixed during a day is expended either in autotrophic respiration or else allocated to one of the three plant tissue pools, Cf, Cw or Cr.
2. Autotrophic respiration is a constant fraction of the C fixed during a day (Waring et al., 1998).
3. Allocation fractions to vegetation pools are donor-controlled functions which have constant rate parameters.
4. For the deciduous model, the timing of initial leaf out is controlled by a simple growing degree day accumulation, and leaf fall by a minimum temperature threshold. The maximum amount of C that can be allocated to leaves is also limited by a parameter (Cfmax)
5. All C losses are via mineralisation (i.e. no dissolved losses).
The aggregated canopy model (ACM) (Williams et al., 1997) is used to calculate daily photosynthesis (GPP) in DALEC. ACM is a ‘big leaf’, daily time- step model that estimates GPP using a simple aggregated set of equations operating on cumulative or average values of leaf area index (LAI, determined directly from Cf, total foliar C), foliar nitrogen, total daily irradiance, minimum and maximum daily temperature, day length, atmospheric CO2 concentration, water potential gradient (cd) and total soil-plant hydraulic resistance (rtot). ACM contains 10 parameters which have been calibrated using a fine-scale model (the soil-plant-atmosphere model (SPA), (Williams et al., 1996) across a wide range of driving variables producing a universal parameter set which maintains the essential behaviour of the fine-scale model but at a much reduced complexity.
You can download the code and drivers for DALEC v1 that were used in the REFLEX project here.
This version produces a single model capable of varied phenological behaviour, and so can be applied globally to forest systems. The model contains six carbon pools (Cfol, Clab, Croo, Cwood, Clilt, Csom). New equations define the time of leaf onset (labile to foliar pool C transfer) and leaf fall (foliar to litter pool C transfer). The new functions are advantageous in that (a) the daily turnover rates result in a continuous and specified loss of carbon throughout a known time period, and (b) the functions are cyclical and hence do not need to be reset, switched on or switched off throughout the model run period, as was the case in v1. In v2 allocation of GPP now proceeds to all living pools (i.e. including Clab), and there is no retranslation of C from leaves to labile pool at senescence.A publication is forthcoming on this version.
Fox, A, M Williams, AD Richardson, D Cameron, JH Gove, T Quaife, D Ricciuto, M Reichstein, E Tomelleri, C Trudinger, MT van Wijk (2009) The REFLEX project: comparing different algorithms and implementations for the inversion of a terrestrial ecosystem model against eddy covariance data, Agricultural and Forest Meteorology 149, 1597–1615.
Williams, M., E.B Rastetter, D.N Fernandes, M.L. Goulden, G.R. Shaver & L.C. Johnson (1997) Predicting gross primary productivity in terrestrial ecosystems. Ecological Applications 7: 882-894.
Williams, M., P.A. Schwarz, B.E. Law, J. Irvine & M. Kurpius (2005). An improved analysis of forest carbon dynamics using data assimilation. Global Change Biology 11, 89-105.