# What kind of output is generated by COSIMA?

At the end of a successfull simulation run COSIMA will generate three output files in simple text format (out.txt, dist.txt, res.txt) which may be inspected immediately in the browser by clicking the tabs "**Detailed overview**", "**Size distributions**", or "**Tabulated results**", respectively.
Furthermore, some selected results are also made available in graphical form in the window "**Selected plots**".

To facilitate the storage of results as well as further processing by the user, e.g. import into plot programs or spreadsheet applications, these output files together with three input files corresponding to the three input boxes (in-parameter.txt, in-leakage.txt, in-distribution.txt) and the automatically generated graphs (combined into one page in png-format) will also be available for download in zip-compressed form.

In case of error conditions, COSIMA will provide all output that has been generated prior to the crash. If error messages are available, these will be displayed in a supplementary tab "**Execution error**". Please bear in mind, however, that incorrect or physically questionable input does not necessarily result in error messages or abnormal program termination. It is the responsibility of the user to provide meaningful input for meaningful results.

#### Detailed overview (out.txt)

In traditional program output form, this file provides a comprehensive summary of the user supplied (or default) input, relevant aerosol characteristics derived from the input data and a time evolution monitor of the simulation run. Because its content is largely self-explanatory to users having at least skimmed the input instruction page, only a few specific details will be outlined below.

In the block "**INPUT DATA AND INITIALISATION MONITOR**" COSIMA offers a table listing for each size bin of the sectional

representation a number of characteristic quantities that determine - within the framework of the employed formalism -

the dynamics of the airborne fractal like particles. These are:

R(mass): mass equivalent particle radius [cm]

R(geo): geometric radius (the radius of the closest spherically symmetric envelope around the fractal like particle) [cm]

R(me,c): mobility equivalent radius in the continuum regime [cm]

R(me): mobility equivalent radius valid for the simulated conditions (typically transition regime) [cm]. Please note that

R(me) and R(me,c) usually differ for fractal like particles outside the continuum regime.

h(K-R): Kirkwood-Riseman factor R(me,c)/R(geo), i.e. a measure for the amount of hydrodynamic interactions and screening

acting on the primary particles inside a fractal cluster. Since h(K-R) ≤ 1, fractal like particles are usually more mobile

than expected from their geometric size.

S(acc): accessible particle surface area, i.e. the fraction of the surface accessible to gas molecules on linear trajectories [cm^{2}].

Under free molecular conditions S(acc) is expected to correspond to 4πR(me)^{2}.

Nprim: number of monomers equivalent to R(mass) (monodisperse distribution of primary particles assumed)

For compact spherical particles R(mass), R(geo), R(me,c) and R(me) are equal, h(K-R) = 1, and S(acc) = 4πR(mass)^{2}.

In this case COSIMA acts like a conventional aerosol behaviour code.

In the section "**SIMULATION RUN MONITOR**" output is provided at time intervals determined by NWRITE(1) and NWRITE(2) (cf. line 3 of the input block "Parameter") and the length of the time integration steps. After NWRITE(1) time steps a short output is generated
comprising CPU and simulation timing information followed by a comprehensive listing of aerosol properties and balances for wall
deposition and leakage induced particle loss. After NWRITE(2) time steps and at the points in time specified for output of the actual size distribution (cf. line 11 of "Parameter" block), this is supplemented by the actual number and mass size distributions (with respect to the mass equivalent radius). The table entries are:

R(mass): mass equivalent radius of section [cm]

dN/dlnR(mass): value of number size distribution with respect to R(mass) [cm^{-3}]

dM/dlnR(mass): value of mass distribution with respect to R(mass) [g/cm^{3}]

Part.-number: number of particles accumulated in section [cm^{-3}]

Mass dist.: mass accumulated in section [g/cm^{3}]

Since only sections with nonzero number concentration are listed, they usually amount to less than K2, the number of bins specified by the user or by default. While this table serves several purposes, it proves particularly useful for controlling the suitability of the chosen size range of the sectional representation. Clearly one wants to avoid a significant cutoff of the size distribution at the upper or lower end, although COSIMA will conserve mass by accumulating particles virtually threatened by loss in the highest and/or lowest bin. However, it is also unfavourable to select the size range in a manner leaving many bins empty over the whole time of the simulation. This way "resolution" is lost resulting in increased numerical diffusion and deteriorated accuracy of the results.

Some shortcuts used within the "**SIMULATION RUN MONITOR**":

Av. radius: average (mean) mass equivalent radius

Rg: median mass equivalent radius

sigma: standard deviation of size distribution related to mass equivalent radius

R50_mass: mass equivalent radius below and above of which the mass distribution comprises of 50% of the total mass,

respectively

Av. me radius: average (mean) mobility equivalent radius

Rg_me: median mobility equivalent radius

sigma_me: standard deviation of mobility related size distribution

Multiplied by the square of the total number concentration, the coagulation constant (Coagulation const.) yields the actual coagulation rate in [cm

^{-3}s

^{-1}].

#### Size distributions (dist.txt)

This file contains the computed number size distributions with respect to mobility equivalent diameter, dN/dln(dme), for the times specified in line 11 of the input block "Parameter". The data are provided in tabulated form appropriate for import into spread sheet based applications:

time_1 time_2 ...... ...... time_m

dme_1 dN/dln(dme)_11 ... ... ... dN/dln(dme)_1m

dme_2 ... ...

...

... ...

dme_n dN/dln(dme)_n1 ... dN/dln(dme)_nm

time_i: simulation time point for ith size distribution [min]

dme_k: mobility equivalent diameter of kth size bin [nm]

dn/dln(dme)_ki: value of discretised number size distribution for bin k at time i [cm^{-3}]

n = K2 (cf. line 3 of input block "Parameter")

m = NOUTP (cf. line 10 of input block "Parameter")

#### Tabulated results (res.txt)

This file provides calculated time evolutions of relevant quantities characterising the aerosol under consideration in tabulated form appropriate for import into spread sheet based applications. The structure of the data set is outlined below:

Column 1: time [h]

Column 2: time [min]

Column 3: number concentration [cm^{-3}]

Column 4: mass concentration [μg/m^{3}]

Column 5: concentration of accessible surface area [cm^{2}/cm^{3}]

Column 6: mass equivalent median diameter [nm]

Column 7: standard deviation of size distribution with respect to mass equivalent diameter

Column 8: mobility equivalent median diameter [nm]

Column 9: standard deviation of mobility related size distribution

Column 10: accumulated diffusional mass deposition [g/cm^{2}]

Column 11: accumulated sedimentational mass deposition [g/cm^{2}]

Column 12: thermophoretic mass deposition [g/cm^{2}]

Column 13: total leaked/sampled mass [g]

Column 14: light absorption [1/m]

Column 15: light scattering [1/m]

Column 16: light extinction [1/m]

#### Selected plots

Graphical output is generated using the Gnuplot software. Currently four graphs are provided:

Plot 1: Time evolution of number and mass concentration (from columns 1, 2, and 3 of file res.txt)

Plot 2: Time evolution of median mobility diameter and standard deviation of mobility size distribution

(from columns 1, 8, and 9 of file res.txt)

Plot 3: Time evolution of mobility related number size distribution dN/dln(d_{me}). A maximum of 20 distributions is displayed

according to subsequent points in time as specified by the user (cf. line 11 of input block "Parameter"). Please note that,

for technical reasons, in this graph the contents of bins containing less than 5/cm^{3} are interpolated from neighbouring

bins.

Plot 4: Time evolution of mobility related number size distribution dN/dln(d_{me}). Up to 300 distributions (typically 10 - 50) are

combined in a smoothly interpolated contour type representation according to subsequent points in time automatically

chosen by COSIMA.

Please note that running very extreme simulation scenarios may result in overflowing the contour plot buffer. In such

cases the generated plot will not cover the whole time range specified by the user.