United States
             Environmental Protection
             Agency
             Environmental Research
             Laboratory
             Corvallis OR 97330
EPA-600/3-78-053
May 1978
             Research and Development
x>EPA
Simulation
of Nutrient  Loss
From Soils
Due to Rainfall
Acidity

-------
                RESEARCH REPORTSNG SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology.  Elimination  of  traditional grouping was consciously
planned to foster technology transfer and a maximum interface in  related fields.
The nine series are:

      1.   Environmental  Health Effects Research
      2.   Environmental  Protection Technology
      3   Ecological  Research
      4.   Environmental  Monitoring
      5.   Socioeconomic Environmental Studies
      6.   Scientific and Technical  Assessment Reports (STAR)
      7   Interagency Energy-Environment Research and Development
      8   "Special" Reports
      9.   Miscellaneous Reports

This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal spe-
cies, and materials. Problems  are assessed for their  long- and short-term influ-
ences  Investigations  include formation, transport, and pathway studies to deter-
mine the fate of pollutants and their effects. This work provides the technical basis
for setting standards to minimize undesirable changes in living organisms in the
aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161

-------
                                          EPA 600/3-78-053
                                          May 1978
     SIMULATION OF NUTRIENT LOSS FROM
       SOILS DUE TO RAINFALL ACIDITY
                    by
               John 0. Reuss
         Colorado State University
       Fort Collins, Colorado  80523
              Project Officer

              Donald J. Lewis
Corvallis Environmental Research Laboratory
         Corvallis, Oregon  97330
CORVALLIS ENVIRONMENTAL RESEARCH LABORATORY
    OFFICE OF RESEARCH AND DEVELOPMENT
   U.S. ENVIRONMENTAL PROTECTION AGENCY
         CORVALLIS, OREGON  97330

-------
                                  DISCLAIMER
     This report has been reviewed by the Corvallis Environmental  Laboratory,
U.S. Environmental  Protection Agency, and approved for publication.   Approval
does not signify that the contents necessarily reflect the views and policies
of the U.S. Environmental Protection Agency, nor does mention of trade names
or commercial  products constitute endorsement or recommendation for  use.

-------
                                   FOREWORD
     Effective regulatory and enforcement actions by the Environmental  Pro-
tection Agency would be virtually impossible without sound scientific  data on
pollutants and their impact on environmental stability and human health.
Responsibility for building this data base has been assigned to EPA's  Office
of Research and Development and its 15 major field installations,  one  of
which is the Corvallis Environmental  Research Laboratory (CERL).

     The primary mission of the Corvallis Laboratory is research on the
effects of environmental pollutants on terrestrial, freshwater, and marine
ecosystems; the behavior, effects and control of pollutants in lake systems;
and the development of predictive models on the movement of pollutants in the
biosphere.

     This report describes a simulation model designed to predict the  impact
of rainfall acidity on the leaching of cations from non-calcareous soils.
This work was undertaken as a part of a research program at CERL to determine
the effects of acid rain on forest ecosystems.


                                             A. F. Bartsch
                                             Director, CERL

-------
                                   ABSTRACT
     This paper describes a simulation model that provides a quantitative
system utilizing established relationships from soil chemistry to predict the
most likely effect of rainfall  acidity on the leaching of cations from non-
c'iicareous soils.

     The model utilizes the relationships between lime potential (pH -
'i/2pCa) and base saturation described by Clark and Hill (Soil  Sci. Soc.
Arner, Proc. 28:490-492, 1962) and Turner and Clark (Soil  $_ci. 99:194-199,
1964), the equilibrium between C02 partial pressure and H  and HC03  in
solution, the apparent solubility product of A1(OH)3, the equilibrium of
cations and anions in solution, the Freundlich isotherm description of sul-
fate adsorption, and mass balance considerations, to predict the distribution
of ions between the solution and sorbed or exchangeable phases.  Ionic compo-
sition of leachates in response to rainfall  c_omposjtion c_an thus be computed.
ions^considered in the present version are H , Ca2 , Al3  , SO^2 , Cl~, and
HC03 .

     The model predicts almost exact chemical equivalence between basic
cc'ciuii removed in the leachate and strong acid anions entering the system in
rh- rainfall   if pH - l/2pCa is above 3.0, at which point the base saturation
wit!  generally not exceed 20|.   At lo^er pH - l/2pCa values leaching of
oinohs in association with H  and Al3  becomes significant and these cations
predominate when pH - l/2pCa falls below 2.0.

     If the soil exhibits sulfate adsorption properties,  leaching of bases in
•expense to rainfall containing sulfuric acid may be delayed until  a sulfate
•:i-!:.0!'ption equilibrium is reached, but base removal would continue after the
jiilfuric acid input was stopped.

     itjst, of the work reported here was completed while the author was on
ciiignrn^nt as a soil chemist at the Corvallis Environmental Research Labora-
tory , Corval1is, Oregon.
                                    IV

-------
                                   CONTENTS


Foreword 	  ...........    Ill

Abstract 	  ............     iv

Figures  	  ..............     vi

Tables	  viii

Acknowledgement  	  ................     ix

     1.   Introduction	      1
     2.   Conclusions  	  ..............      2
     3.   Recommendations  	  .............      4
     4.   Model Structure and Theory 	  .......    .  .      5
     5.   Simulation Results ...  	  ........       17
     6.   Discussion 	  ...................     35
References

-------
                                LIST OF FIGURES
Number                                                                Page

 1.  Generalized diagram of model  describing  interaction of rainfall
     chemistry with soil  solution  processes	    6

 2.  The pH - l/2p(Ca  + Mg) values of soils as  a  function of base
     saturation based  on the sum of cations extracted with IN KC1.
     (Redrawn from Clark and Hill  1964)	10

 3.  Flow diagram for  model of the rainfall chemistry-soil solution
     system	15

 4.  Model  prediction  of pH of a 1:1  soil:water extract in a soil  with
     pH - 1/2 pCa = 3.0 as a function+of [SO^2" + Cl~] concentration  in
     rainfall.  Circles represent  Ca2 as rainfall  while X's represent
     pH 4.0 rain as shown in Table 1. The soil does  not adsorb SO^2"
     and leachate volume is equal  to rainfall	19

 5.  Model  prediction  of_concentration of Ca2   in leachate as a func-
     tion of [SO^2' +  Cl~] concentration in rainfall  and pH - l/2pCa
     of the soil.  The soil does not adsorb SO^2' and leachate
     volume is equal  to rainfall	21

 6.  Ratio of Ca2  in  leachate per [SOit2" + Cl~]  in rainfall (eq/eq),
     as a function of  anion strength of  rainfall  and  pH - l/2pCa.   Soil
     does not adsorb sulfate	22

 7.  Predicted effect  of Changing  evaporation  from 0  to 90% of rainfall
     on the amount of  Ca2  leached from  the soil  per  equivalent of rain.
     pH - l/2pCa was 2.5, rainfall pH 4.0, and  there  was no adsorption
     of S042~	25

 8.  Model  prediction  of concentrations  of Ca2   in leachate of a
     S042" adsorbing soil  as a function of SO^2' + Cl" and pH-
     l/2pCa.  Leachate volume is equal to rainfall	27

 9.  Simulated time-based fluctuations in soil  pH using pH 4.0 rain-
     fall	31

10.  Simulated change  in total Ca2  (soluble  plus exchangeable) in
     0.3 meters of soil during a 5-year  peiod  using pH 4.0 rainfall.  .   33
                                    VI

-------
"11.  Simulated total  Ca2  found on April  30 in 0.3 meters of soil
     stressed with pH 4.0 rainfall over a 10-year period. ...  	   34

12.  Simulated April  30 1:1  soiltwater extract pH values for SO^2"
     adsorbing and non-adsorbing soils stressed with pH 4.0 rain-
     fall  	  .........  	   35

-------
                                LIST OF TABLES


Number                                                           Page

 1.  Ion concentrations In rainfall  used in similation runs.  .  .    18

 2.  Effects of C02 partial  pressure on extract pH and on
     Ca2  leached using simulated pH 4.0 rainfall  	    24

 3.  Initial conditions and soil  properties for a  10-year
     acid rainfall  simulation	     29

 4.  Rainfall and evapotranspiration patterns used for 10-year
     simulations	     30
                                    vm

-------
                               ACKNOWLEDGEMENTS
     An undertaking such as that described in this paper requires the contri
butions of many people.  I would like to express my appreciation to the
administration and staff of the Corvallis Environmental  Research Laboratory
who made my assignment at the laboratory possible, and provided facilities
and encouragement for the completion of the task, particularly Dr.  Norman
Glass, Dr. Allen Le Fohn, and Dr. Larry Raniere.  Thanks are also due to Fir.
Donald Lewis, who helped in the final stages of the work.

     My thanks for their efforts and suggestions are due those who  reviewed
the manuscript including Dr. C.V. Ccle (USDA-ARS), Dr. Willard Lindsay, and
Dr. Robert Woodmansee, all from Colorado State University, and Dr.  Dale
Johnson, Oak Ridge National Laboratory.

     An extra measure of thanks are due to Mr. J. Warren Hart who's efforts
and skills in programming, mathematics, and chemistry contributed substan-
tially to the successful planning of the program.

-------
                                  SECTION 1
                                INTRODUCTION
    The trend toward increased rainfall  acidity over northern and western
Europe and the northeastern United States appears well  established.   This in
turn is causing concern about the possibility of accelerated losses  of miner-
al  bases from soils and subsequent losses of productivity (Anon., 1971; Jons-
son and Sundberg, 1972).  However, ion exchange properties in soils  are com-
plex, and cation losses from soils due to leaching by acid rainfall  is modi-
fied by soil  properties (Wiklander, 1974; Malmer and Nilsson, 1972).
    This paper describes a model that calculates ion loss from soils  as a
function of soil properties and of the composition and  distribution  of rain-
fall.  The objective of the model is to  provide a quantitative system that
utilizes established principles of soil  chemistry to predict the most likely
effect of rainfall  acidity on leaching of basic cations from non-calcareous
soils.  It must be regarded as preliminary due to the limited number of ions
considered and numerous simplifying assumptions.  Nonetheless, in my opinion,
the model provides the most reliable method presently available for  estima-
ting the effect of rainfall composition  on losses of bases from soils.  It
also provides the basic structure that can be used in subsequent models inco,
porating additional ions and processes.

-------
                                   SECTION  2
                                  CONCLUSIONS
     It is entirely feasible to simulate the  ion exchange and removal  pro-
cesses that are likely to occur in soils as a  consequence of rainfall  acidity.
Known physical-chemical  relationships  coupled  with  simulation techniques  com-
bine to give the best prediction presently  possible of long  term effects  that
are not amenable to experimental measurements  within a reasonable length  of
time.  The following specific conclusions were arrived at by a first generation
model of this type that applies only to non-calcareous soils.
     The initial results of these investigations indicate that significant
acidification and depletion of bases could  occur over a period of a few de-
cades if rainfall inputs are consistently acidified to pH 4.0 with sulfuric
acid.
     He can expect almost an exact chemical equivalency between strong acid an-
ion content of rainfall  and leaching of bases  from  soils with base saturation
levels above about 20%.   As base saturation becomes lower the leaching will
diminish and then cease but the exact nature  of this relationship must await
further investigation.
     Leachate composition will initially be controlled largely by the anion
concentration of the rainfall, and will be  independent of whether the* rainfall
cations are mineral bases or H .  The long  term base status  of the soils will,
however, respond to the cation composition  of the rainfall.
     Soils with a significant capacity to adsorb sulfates will tend to dampen
the effect of sulfuric acid induced leaching  of basic cations.  The leaching
of bases will respond more slowly to acidic imputs  than on non-sulfate adsorb-
ing soils but will reach a similar equilibrium rate of leaching as the sulfate
adsorption capacity reaches equilibrium with the rainfall sulfate.  On these

-------
soils high rates of leaching induced by rainfall containing sulfuric acid
eoiild be expected to persist for some time after acid rainfall input ceased.
    Soil pH as measured by soil water suspensions may be markedly lowered by
increased rainfall ion concentrations.  This rapid change is a result of rain-
fall anion concentration and would occur in a similar manner in response to
rainfall containing either strong acids or neutral salt.
    Soil pH as measured by soil water suspension is well known to vary with
time.  However, the predictions given by the model show clearly the effect of
seasonal precipitation-evaporation relationships of a magnitude of about 2.5
times the overall change that would be expected to occur in a decade in res-
ponse to pH 4.0 rainfall.  Experiments that depend on soil  water extract or
suspension pH measurements to evaluate the effect of acid rainfall inputs
would be subject to serious errors due to these fluctuations.
    The long term leaching of cations in response to rainfall acidity will be
largely independent of the precipitation/evaporation ratio as long as periodic
leaching occurs.
    Cation removal in response to rainfall mineral composition will be largely
independent of CQ? partial pressure.
    The initial model results indicate that base removal due to rainfall acid-
ity from actual soil systems can validly be investigated on an experimental
basis by increasing the acidity of applied water to levels in excess of that
normally found in rain.  This would shorten the time necessary to investigate
base removal.  Measurement of pH on soil water suspensions from such experi-
ments would not be valid as much lower values would be expected than would oc-
cur if  the weaker acid were applied for the longer time periods.

-------
                                  SECTION 3
                               RECOMMENDATIONS
    Simulation modelling based on known physical-chemical  relationships in
soil offers the most feasible method of evaluating long term effects of acid
rainfall  on the base status of soils.  The model  reported  here is useful  in
this regard but is limited by the inclusion of only part of the important ions
and by simplifying assumptions concerning water movement in soils.   The rela-
tionships are sufficiently understood to develop a more complete model  inclu-
ding additional ions and more sophisticated water movement relationships.
Such a model  should be developed immediately.
    At the same time laboratory work with soil suspensions and soil  columns
should be undertaken for experimental verification of the  model.  If results
coincide  with model predictions it will add confidence to  model prediction of
long term effects not easily amenable to direct experimentation.  If predic-
tions of  laboratory systems do not give acceptable precision,  hopefully the
reason can be discerned and the model modified accordingly.
    The model should also be used to simulate  on-going field experimentation,
both to validate the model, and to determine if the procedure  used  in the
field trials  can reasonably be expected to give reliable and useful  results.

-------
                                   SECTION 4
                          MODEL STRUCTURE AND THEORY
GENERAL DESCRIPTION
     A general schematic representation'of the model is shown in Fig. 1.   The
model was developed using a daily time step.  Data supplied on a daily basis
from external sources include rainfall and evapotranspiration in millimeters,
                           2+   +    -         2-
and the concentration of Ca  , H , Cl , and SO.   ions (moles/liter) in the
rain.  The present form of the model has been restricted to these ions plus
HC03 , and Al  .  Various initial conditions must be specified at the start of
each run and these will be pointed out in the discussion of the soil process-
es.
     After the rain has entered the soil composition of the soil water is recal-
culated based on previous water content, constituents added, and evapotranspir-
ation.  The ionic equilibria are then calculated and a new solution concentra-
tion is established for each ion.  If the new water content is greater than
field capacity, the excess is removed as leachate, which is assumed to have
                                                                   2+     2-
the same composition as the soil solution.  The total amounts of Ca  , SO.  ,
and Cl" ions in the system are specified initially.  The amount entering  in
the rain is known from the input data and the amount lost in the leachate at
each time step are calculated within the model, so the total sorbed plus  sol-
ution amount of each ion is known at all times.
EQUILIBRIUM PROCESSES
     The equilibrium between solution ions and sorbed ions is the principle fo-
cus of the model.  In the initial stages it was assumed that the soil-solution
and atmosphere-solution equilibria were uniform to the depth being considered
and all results reported are from this version.  Later revisions include  the
capabilities of considering various depth layers with different soil properties
separately.
                                     5

-------
Soil Surface
                   Evapotranspi ration
                          A
                           Rain  FLO

                           Ca2+, H+

                       HCO  ~   Cl~, SO
                                                                   2-
    C02<:
                   Soil  Solution

                =>  A13+, Ca2+,  H+
HCO,
Bottom of Rooting Zone
,  S0
     2-
                 Sorption
                 Processes
    Sorbed
     Ions
Ca
                                                                 ++
        -, SO
                                                                            2-
                                  Leachate
Figure 1.   Generalized diagram of model  describing the interaction of rainfall

           chemistry with soil solution  processes.

-------
Sulfate Adsorption
    Soils exhibit the phenomenon known as sulfate adsorption to varying de-
grees.  Sulfate adsorption is commonly encountered in acid soils, particularly
those that are highly weathered and contain iron and aluminum sesquioxides.
Haward and Reisenauer (1965) discuss possible mechanisms of sulfate adsorption
processes.  Whatever the mechanism, soils that adsorb sulfate tend to maintain
                          2_
an equilibrium between SO.   in solution and that sorbed onto soil surfaces.
                     0                         O
                  S04   (solution) <      > S04   (sorbed)

Various expressions have been used to describe this equilibrium, including the
Freundlich and Langmuir adsorption isotherms (Chao e_t aj_. , 1962; Hasan et al . ,
1970)
                                     2_
    I have chosen to describe the SO.   adsorption process in the model with
the Langmuir equation which is used in the form:
                           ,             K    [SO.2"]
                        SO/' (sorbed) = -™* - 1—                 (1)
          2_
Where [SO.  ]  is the equilibrium sulfate concentration in moles/liter,  K
         M-                                o                             IMaX
is the adsorption maximum, i.e., moles SO. ~ adsorbed per unit of soil  at in-
           2-                                                          2-
finite [SO.  ], and k:  is related to the bonding energy and is the [SO    ]  at
         2 •          ^                                   2-4
which SO.  adsorbed is equal  to one-half K   .   Total S04   is always known
and consists of the sum of the sorbed and solution components so that:

                 S042" (total) = S042" (sorbed) + SO/' 0 D           (2)

Where 0 is the volumetric moisture content (liters water/liter soil)  and D is
                                                                         2_
the depth in millimeters of the soil profile simulated.  The units of SO.
              2                                                         ^
sorbed and S04   total as used in the model are moles per square meter  to
depth D.   Equations (1) and (2) are sufficient to define the distribution of
total sulfate between the solution and sorbed phases, and this distribution is
calculated in the model at each time step prior to calculating the concentra-
tion of the remaining ions as described below.
    These calculations depend on the availability of values of K    and kj  for
                                                                max      -^
any soils to be simulated.  Many soils adsorb very little sulfate and can be

-------
 simulated by simply inserting very low values for K    .  Literature values are
                                                   max
 available for a number of typical sulfate adsorbing soils  (Chao et_ al_.,  1962;
 Hasan et al_., 1970).
     Chao e^t a]_. (1962) point out the limitations of the Langmuir  isotherm.
 The soils they studied did not exhibit appropriate adsorption maximas, even
 though they did follow the relationship over a substantial range from which an
 "adsorption maxima" could be predicted.  Hasan ert a]_.  (1970) refer to this as
 a  "first phase adsorption maxima."  Thus for any particular soil simulated, the
 relationships in the model are only valid for the range of sulfate adsorption
 within which the Langmuir relationship is valid.  In addition it should  be un-
                     2_
 derstood that the SO.   concentration in (1) might better be expressed in
 terms of activity.  However, the literature values available are generally in
                                                                      2_
 terms of concentration so no activity coefficients are  applied to SO.   in
 the model.
 Ca - H Equilibria
                             +       2+
     In order to define the H  and Ca   exchange equilibria, I chose to  util-
 ize the concept of "Lime Potential" of Schofield and Taylor (1955).  Soil pH
 measurements are highly sensitive to factors such as dilution and salt effects,
 but these authors demonstrated that the Lime Potential, defined as
 pH - %p(Ca  + Mg), is remarkably constant for a given soil even though soil wa-
 ter content may vary widely.  With changes in water content the shifts between
                            2+      +
 exchangeable and solution Ca   and H  are such that K.  remains constant.  In
 our system  where Mg is not presently considered we can write:
                               pH - %pCa = KL                          (3)

 K, , the Lime Potential, can easily be determined for any given soil by measur-
                                     2+
 ing the pH  and the activity of the Ca   ion in a suspension of soil.   pH and
                                                                  +       2+
 pCa are defined as the negative logarithms to the base 10 of the H  and Ca
activities  respectively.
     The relationship in  (3) is sufficient for our purposes if only short term
simulations are used.   However, if losses or gains of calcium are significant
                                             2+
the fraction of exchange sites occupied by Ca   (base saturation) changes so
that KL changes also.   It appears, however, that K,  can be satisfactorily rel-

-------
ated to base saturation as shown by Clark and Hill (1964).  In our system, the
only base considered is calcium, so base saturation V is defined by
                                 V = 2CaX
                                     CEC
Where  CaX  is the moles of exchangeable Ca per unit of soil and CEC is the
cation exchange capacity in equivalents.  If we accept the relationships found
by Clark and Hill (1964), (Fig. 2) one for montmorillonitic soils, and the
other for more weathered soils, we can write:
                              PH - hpCa = f()                      (5)
        CaX
Where ^FFK) is the function shown in Fig. 2 for the soil  we wish to simulate.
In the program coordinates of the points on the appropriate curves are fur-
nished as data and intermediate points are calculated by linear interpolation
as required.
    Some care is required in the interpretation of Fig.  2  due to the manner
in which CEC is defined.  Clark and Hill refer to this as  "permanent", i.e.
non-pH dependent, charge.  In reality in highly base saturated soils, some pH
dependent charge would be included.  Thus, if a highly base saturated soil is
acidified, CEC may decrease.  Simulations resulting in large changes in base
saturation could be subject to error from this effect.
                                                               2+
    In addition to the relationships shown in (5), the total Ca   in the simu-
lated system is always known, so that:

                         Catotal = CaX + [Ca2+]  G D                   (6)
        2+
Where Ca   is the solution concentration in moles/liter, and CaX is the ex-
                        2
changeable Ca in moles/in  to whatever depth is considered, and D is expressed
in millimeters.
Chemical Equilibria
    In addition to the above, certain chemical conditions  must be satisfied.
In soils above pH 5.0 the HCO ~ ion may be an important constituent.  The con-
                             O
centration of this ion is controlled by C02 partial pressure such that:

-------

6.0
     5.0
     4.0
      3.0
     2.0
                 •  Montmorillonitic
                 o  Others
               20      40      60      80
                    %  Base Saturation
                                         100
Figure 2.   The  pH  -  ^plCa + Mg) values of soils as a
           function  of  base saturation based on the
           sum  of  cations extracted with IN KC1.
           (Redrawn  from Clark and Hill, 1964).
                           10

-------
                        (H+) (HCO ") = (COJ  • lO"7'81                 (7)
                                 O        C-
Where (C02)  is the partial pressure of CCL in atmospheres, and (H ) and (HC03~)
are the activities of the two ions in moles/liter.  In the model the CCL par-
tial pressure must be specified initially.
     In acid soils Al   may be one of the major cations in the system, and neg-
                                                                  +       2+
lecting this ion results in unacceptable errors in the predicted H  and Ca
solution concentrations.  From the solubility product of A1(OH)3 we can write:

                             (A13+) (OH")3 =  KA1
                 10-14
and substituting - r-  for (OH) we obtain eq. (8).
                  (H+)

                                     -14 3
                           (A13+) (-^-f-)  =  Kfl1                       (8)
                                    (H+)      A1
     In soils the K», would represent the apparent solubility product for
Al(OH)o which may vary depending on the degree of weathering.  For the system
                            _oo /i      -T3 R
simulated here, values of 10"  '  to 10"  '  were used (Turner and Clark, 1964;
Turner, Nichol and Brydon, 1963).
     In order to maintain electrical neutrality in the system the total equiva-
lents of anions and cations in the solution must be equal, so that:

           [H+] + 2[Ca2+]  + 3[A13+]  = 2[S02"] + [HC0~]  + [CM       (9)
The brackets again indicate molar concentrations.  In this system Cl  is pre-
sumed to exist only in solution, so that when the initial amount is specified
and the gain in rainfall and losses in leachate are all known, the amount pre-
sent at each time step can be calculated.
     If we examine eqs. (5) through (9), we find the following parameters are
either known or specified; CEC, f (-) , Catotal ' 03 D' ^Vg' tS042"] and
[Cl"].  The unknowns are (Ca2+), (H+), [Ca2+] , CaX, (HCO "), [HCO ~] ,  [H+]
       3+                                               <5        3
and (Al  ).  If for the moment we consider activities (   ) to be equal to con-
centrations [    ] , we have only [Ca  ] ,  [Al +] ,  [H+] ,  [HC03~] , and CaX as

                                    11

-------
unknowns.
    The five equations in five unknowns define the system, so by solving this
set we could calculate the solution concentrations of Ca   , H , Al  , HCCL ,
                       2+
and the exchangeable Ca  .
    An assumption that activities and concentrations are equal is not accept-
                                                   3+
able, particularly in relation to the tri-valent Al   ion.  It is necessary,
therefore, to calculate the activity coefficients y., Y2> Y3» where:
                       Y, [Concentration]  = (activity)               (10)
The Debye-Huckel Theory is used to calculate these coefficients in a manner
similar to that utilized by Dutt et al_., (1972).

                                      -.509Z.V5
                             log Y, = 	V~                     (ID
                                  1     1  + y'5
Where Z is the valence and:
                                   n       0
                             y = h I ,  C.r.                         (12)
                                   i=l   i  i
Where n is the total number of ion species  present, and C. is the concentra-
tion of each ion.
    Equations (1) through (12) are sufficient to define the solution concen-
trations of the ions considered as well as  the amounts sorbed provided the to-
tal amounts of input and output are known.   Thus by devising a solution to
this set of relationships it is possible to predict the changes that will take
place as a result of any rainfall composition, provided the amount and distri-
bution of rainfall and the evapotranspirations are known.
    Some cautiqn is necessary in interpreting ion concentration or activities
as calculated from these equations, particularly in terms of (H ) or ph'.  The
ion activities reported here as calculated  from the above equations should
represent those found in the solution phase, commonly referred to as the "outer
solution."  The boundary between "inner solution" or sorbed phase and "outer
solution!!may not be well  defined and pH measurements made on soil suspensions
or pastes are generally lower than those of the extract or supernatant solu-
tion due to the influence of this "inner"  phase.  Probably the best interpre-

                                     12

-------
tation of the pH values calculated by the program would be those of an extract
from a paste or suspension prepared from the indicated soi1:water ratio.
MODEL STRUCTURE
Numerical Solutions
    The heart of the model is the numerical  solution of sets of equations  des-
cribing sorption and chemical equilibria.  The system utilized evolved during
development of the model and is probably not the most efficient that could be
devised.  However, it works well and the processor time required is moderate,
even for long term simulations.  A brief description of the approach seems
appropriate.
    First the activity of Al   is calculated from eq. (8)  and the value of
(H ) is taken from the previous time step or iteration.  If no previous value
     +                                                                2-
of (H ) is available, a value is assumed.  The solution and sorbed SO.  are
then calculated using eq. (1) and (2).   Activity coefficients are calculated
using eq. (11) and (12) and previous concentrations or if none are available
activity coefficients are assumed for the first iteration.
    Next, an iterative method is used for solving equations (5), (6), (7), and
(9).  This solution is the most complicated in the model.   Equations were  suc-
cessively eliminated by substitution and rearranged until  one complicated  ex-
pression of the form f(H ) =0 was obtained.  This is solved by an adapta-
tion of Newton's method.  The value of f(H ) for an assumed value of (H )  is
calculated and the derivative of f(H )  is evaluated.  A new estimate of (H )
is obtained by projecting the tangent to the abcissa.  The processes are re-
peated until the change in the value of (H ) is within limits prescribed by
the operator.   In our simulation this change must be <10   times the new value
of H .  In the initial equilibration, the program assumes a trial value of
(H ), but in subsequent equilibrations the value from the previous time step
is used.  Special precautions are taken to avoid estimates for which f(H)  is
nonexistent and to avoid divergence.  When convergence is satisfactory, new
                                                                        2+
values are calculated for CaX and the concentration and activities of Ca    and
                                                          3+2
HC03 .  The program then returns to the calculation of (Al  ), [SO. ~] and
activity coefficients.  The process is repeated until the changes in the acti-
vity coefficients between successive iterations are within prescribed limits.
                                   13

-------
General Structure
    A generalized flow diagram is shown in Fig.  3.   Some of the steps shown in
this diagram, particularly the equilibrium calculations discussed above, are
quite complex and have their own internal flow patterns.
    First the initial  conditions are read.  These include; constants for
   o
SO.   adsorption, Kfll , depth, CEC, initial water content, field capacity,
  H                ML                        2+               -
permanent wilting points, total  SO., total Ca  , and total Cl   in the soil,
and C0? partial pressure.  Coordinates of points for the pH -  ^pCa as a func-
tion of base saturation relationship are also read  as data, as are the limits
allowed for the checks of convergence.
    The initial equilibria are then calculated for  any five specified water
contents, and the ion concentrations printed.  The  initial ion concentration
at field capacity defines the initial concentration in the leachate.
    If actual rainfall simulations are desired,  rainfall is added.  Rainfall
of any selected composition is presently added in a preset pattern but actual
rainfall and evaporation could be read from a data  file.  After daily rain-
fall is added, evapotranspiration is subracted unless this would reduce soil
water content below the permanent wilting point  in  which case  soil water con-
tent is set at the permanent wilting point.
    The program provides for printed output at preset intervals.  If the date
is one for which printed output is required, the procedure is  as follows:
First, a check is performed to determine whether total water is above field
capacity, and if not, no leaching will occur and the program skips directly
to the output section.  If leaching will occur,  the ion equilibration routines
are utilized to determine solution concentration.  Water in excess ofc field
capacity is removed, and the amounts of ions removed are accumulated.  The
program then moves to the output routines in which  the equilibration routines
are called for the preset output water contents; and the concentrations, the
accumulated amounts of ions added and leached, and  the present amounts of sor-
bed ions are printed.  The program then updates  base saturation, checks for
end of run and returns to add the next day's rain.
    On days when no print is required, the system is simpler;  if water content
is below field capacity, no further calculation  is  required and the program
                                    14

-------
 f   START   J
     /READ
    INITIAL
  CONDITIONS/
Figure 3.  Flow  diagram for model of the  rainfall  chemistry-soil  solu-
           tion  system.
                                   15

-------
 returns to check for end of run and rainfall addition.  If field capacity is
 exceeded, new ion equilibria are calculated, the solution in excess of field
 capacity is removed, and the amounts of water and ions lost are accumulated.
 Base saturation is updated, end of run is checked and the next days rain is
 added.
 Output
     As mentioned previously, output is printed at pre-selected intervals.   For
 the output to be comparable from one time to the next, it is necessary for the
 output equilibria to be calculated for the same water content each time.   Also,
 water contents at which laboratory measurements of pH or solution ion concen-
 tration are made may vary.   For this reason, provison is made for output of
 ion concentrations and sorbed ions at five specified water contents.  The vol-
 umetric moisture contents (9) desired are entered.   In these runs, I have used
 field capacity, approximate saturation (2 x field capacity); and 1:1, 1:5 and
 1:10 soil:water ratios.  The latter was obtained by setting 0 such that bulk
 density/0 is equal  to the desired ratio.
     In addition to the ion concentrations and sorbed ions,  the accumulated
water and the ions  added and leached since the start of the run and during  the
 preceding interval  are printed.
Coding
     The model  is  coded in  the standard FORTRAN IV  computer language.   For  the
version used in the runs reported here, core memory requirement was about
70000 (octal).   Execution time is minimal  and by placing a compiled version on
a system file,  processor charges on the system utilized were less than $2 for
a ten year  simulation run.
                                    16

-------
                                  SECTION 5
                             SIMULATION RESULTS
SHORT-TERM SIMULATIONS
    One of the most important applications of the model  is for short-term
runs in which the ion input-output relationships can be  examined as a fun-
ction of soil and rainfall characteristics.  These simulations provide a ba-
sis for understanding how cation loss due to acid rainfall is influenced by
soil properties.
Non-Sulfate Adsorbing Systems
    First, we shall examine results of short-term simulations of a soil  sys-
tem with very low sulfate adsorbing capacity.  These runs were made simulating
                                                           -4
rainfall varying in pH from 3.00 to 6.35 with C0? at 3 x 10   atmospheres.
The ionic composition of the rain is shown in Table 1.  The pH - %pCa values
of the soil were varied from 1.5 to 4.0, corresponding to soils that are near-
ly base saturated (4.0) and extending below the levels normally expected in
soils.  The pH - ^pCa values were not allowed to change  during a run.  No ev-
aporation was allowed and steady state conditions were required, i.e., the
   2-
SO.   and Cl  concentrations were in equilibrium with those of the rainfall
and did not change during the run.
    Under these conditions, soil solution pH was markedly affected by chang-
ing composition of the rain.  The results for pH - %pCa = 3.0 are shown in
Fig. 4.  This should not be interpreted as due to rainfall acidity but rather
to a "salt effect".  This can easily be demonstrated with the model by repla-
                                                              2+
cing the "acid rainfall" with a CaSO. rainfall in which the Ca   concentra-
                                            2-
tion is adjusted to equal the sum of the SO.   plus Cl  (in equivalents per
liter).  Such rainfall would be essentially neutral but the effect on soil pH
is almost identical to that of acid rainfall of the same anion concentration,,
A conceptual model of this effect is that an increase in anion concentration
                                    17

-------
TABLE 1.  Ion concentrations in rainfall  used in simulation runs.
No.

1.
2.
3.
4.
5.
6.
7.
*
pH

3.00
3.50
4.00
4.49
4.98
5.67
6.35
Ca2*
micro
30
30
30
30
30
30
30
<•
equivalents/liter
1010
330
110
42
20
10
0
Cl

40
40
40
40
40
40
40
log [SO/' + Cl"]
-log (eq/ liter)
2.99
3.46
3.89
4.21
4.40
4.52
4.70
             _
 pH at 3 x 10   atm (XL

-------
              6.0



           Q>



           5



           •I  5.0

           w
              4.0
                  5-0
                                ^   Acid  Rain (pH 4.0)


                                0   Co S04 Rain
              4.5     4.0     3.5      3.0


             log  [SO^'+Cf]  in Roinfdl.eq /liter
Figure 4.  Model prediction of pH of a 1:1 soil: water extract


           in a soil with pH - %pCa = 3.0 as a function of

               2-
           [SO.   + Cl J concentration in rainfall.  Circles

                       2+
           represent Ca   as rainfall while X's represent pH


           4.0 rain as shown in Table 1. The soil does not

                     2_
adsorb SOZ


fall.
                        and leachate volume is equal to rain-
                                19

-------
 in  solution  requires  a corresponding increase in cation concentration, inclu-
 ding  H  .  A  soil  solution with very low anion concentration could not be high-
 ly  acidic.
    Another  characteristic observed in these short-term simulations was that
                       2+
 the concentration of  Ca   in the leachate was strictly dependent on the
    2-
 SCL   + Cl   concentration of the rainfall regardless of whether the rainfall
  4          2+     +                      2+
 cation was Ca   or H  .  The net change in Ca   in the soil, of course, was
                                      2+      +
 markedly affected by  the balance of Ca   and H  in the rain.
                                            2+
    Fig. 5 shows a plot of the log of the Ca   concentration in the leachate
 as  a function of the  log of the anion concentration of the rainfall. Note that
                          2+
 at  pH - igpCa = 3.0, the Ca   concentration of the leachate is almost identi-
 cal to the rainfall anion concentration over the range of 10" '  to 10
 equivalents per liter.  At pH   %pCa above 3.0 and rainfall concentrations be-
      -4 0                                      2++
 low 10  '  equivalents per liter the leachate Ca    concentration is greater
                       2-
 than the sum of the SO.   + Cl  due to significant amounts of HCCL  .  As
                                       2+
 pH  - ^pCa drops to 2.0, the leachate Ca   is substantially lower than the rain-
fall anion concentration, indicating that H  and Al   are significant compo-
nents of the leachate.
                          2+
    The actual ratio of Ca   leached per strong acid anion input in the rain
(eq/eq) is shown in Fig. 6.  Here, values greater than 1.0 indicate leaching
in association with HCO-  while values less than 1.0 indicate strong acid
                                    +       3+
anions leached in association with H  and Al  .   Note that the basic cations
removed are approximately equal to anion input at pH - ^pCa =3.0.  At rain-
fall ion concentrations relevant to the acid rain system (greater than about
10     eq. /liter) even at pH - J^pCa = 2.5 the Ca   leached amounts to about
                              2+
80% of the anion input, but Ca   leaching drops off very rapidly as pH - %pCa
is lowered to 2.0.   A comparison of these values with the curves in Fig. 2 in-
dicates that we would expect almost a stochiometric 1:1 removal  of bases
except on soils of very low base saturation, perhaps less than 20% of perma-
                                                                     2+
nent charge.  However, as base saturation is further depleted, the Ca   remov-
al would decrease rapidly and for the pH 4.00 rainfall  shown in  Table 1, the
  ?+
Ca   input would equal the amount leached at about pH   %pCa = 1.9.
    An important consideration is the effect of evapotranspiration.  If evapo-
                                   20

-------
£4
o
o
Q»  5
O

O 6
O>
O
                     pH-l/2pCo
        5.0
                            4.5
                 4.0      3.5      3.0


-log  [so 4 Cl ]  in Rainfoll, eq/liter
                                        2+
                                           in leachate
Figure 5.  Model  prediction of concentration of  Ca
                                 2-
           as a  function of [SO,    + Cl  ]  concentration in

           rainfall  and  pH  - %pCa of the soil.   The  soil  does
                          2_
           not adsorb  SO    and leachate volume  is. equal  to

           rainfal1.
                     21

-------
            CT
            0)
            CT
            0>
            c
            o
              2.0
            = I  5
            o
            c
            o
            Q:
              0.5J
            o
            o
              0-0
                        2.5
             2.0


             1.9
2.0


1.0
Figure 6.
          5.0      4.5      4.0      55     3.0

         -log  [804" Cl~] in Roin,eq/liter



           2+                      2-
Ratio of Ca    in  leachate  per  [SO.   + Cl ]  in rain-

fall (eq/eq),  as  a  function of anion strength of

rainfall and pH - ^pCa.  Soil  does not adsorb sulfate
                                22

-------
transpiration is always greater than rainfall, removal  is zero and the system
accumulates ions from the rainfall.   In the previous simulation,  evapotrans-
piration was assumed to be zero.   When evaporation occurs the effect is to
concentrate the soil solution and decrease the amount of leachate.  The impli-
cations of this concentrating effect on base leaching were demonstrated by
                                       2
simulating systems in which 10 liters/m  of rainfall per day were applied
with no evaporation.  After equilibrium the evaporation was changed to 9 lit-
ers per day; resulting in leaching dropping from 10 to 1 liters per day-  In
these runs base saturation was held constant to avoid confounding the effect
                                                     2+
of evaporation.  When evaporation was started, the Ca   leached per unit of
rain immediately decreased due to the decrease in amount of leachate (Fig. 7).
The soil leachate than started to become more concentrated until  a new equil-
ibrium with soil solution was established.  In the case shown (pH - %pCa =
2.5), field capacity = 0.30 and depth = 300 millimeters.  This new equili-
                                                          2+
brium was reached in about 1000 days, at which time the Ca   leached per unit
of rain was only very slightly higher (10.34 x 10"  vs. 10.22 x 10"  eq/liter)
than in the system where no evaporation was allowed.  If evaporation was de-
creased instead of increased, a period of increased base loss per unit of
rainfall was obtained, and the resultant plot is virtually the inverse of that
shown in Fig. 7.
    Only a few experiments were conducted with the model in which partial
pressure of C09 was a variable.  Results of one of these (Table 2) show that
                                           -4          -3
at pH - %pCa = 3.0 changing C09 from 3 x 10   to 3 x 10   atmospheres had
                                                  2+
very little effect on soil pH or on leaching of Ca" .  However, at pH - ^pCa
                            2+
=4.0 and 4.5 leaching of Ca   increased by 33 and 78%, respectively, with
                                              2+
the increase in C09.  Increased leaching of Ca   leached due to C09 was ac-
companied by an equivalent increase in HCO ~ in the leachate.  Apparently
             2+
changes in Ca   leaching due to the strong acid anions in the rainfall are
virtually independent of CO- partial pressure.  Leaching of cations in asso-
ciation with HCOZ seems to be relatively unimportant when pH - %pCa is 3.5 or
less but at pH - J^pCa levels above 4.0 it becomes the major factor, particu-
larly if the rainfall is non-acidic.
Sulfate Adsorbing System
    The effects of sulfate adsorption on base removal by acid rainfall as pre-

                                    23

-------
                                                                2+
Table 2.  Effect of CCL partial  pressure  on  extract  pH  and  on  Ca    leached

          using simulated pH  4.0 rainfall.
                  C02                             _   Ca  leached
                                                        o
pH - %pCa        (atm)             pH,  1:1          meq/m      % change
   ~ n          3 x 10":             5.41            12.28
   <3'u          3 x TO'3             5.36            12.80

   . ,          3 x 10"J             5.88            12.94
                3 x 10"°             5.77            14.50

   , n          3 x 10"^             6.34            13.48
   ^'U          3 x TO"3             6.16            17.94

   45          3 x 10-^             6.77            14.72
                3 x 10-3             6.52            26.20
                                    24

-------
         12
      in
       O
       x 10
       0!
       cr  8
       cr
       
-------
dieted by the model are interesting and important.  I know of no experimental
results that either support or refute these predictions.  Therefore the pre-
dictions should be considered as testable hypotheses.
    For these short-term simulations involving a soil with significant sulfate
                                                           _2
adsorption capacity an adsorption maxima (1C,..) of 3.6 x 10   moles/kg and a
                                                /L
one-half maximum concentration (kj) of 1.8 x 10"  moles/liter were used.
                                 -s
These values were calculated from the results of Chao ejt al_. (1962) and app-
                                                     2-
roximate their values for an Astoria soil.   Total SO.   initially present in
                                    -4
the soil was assumed to be 3.15 x 10   moles/kg.
    When short-term simulations were conducted with this system using various
characteristics and values of pH - hpCa ranging from 1.5 to 4.0 (Fig. 8), the
 results were quite different from those of the non-adsorbing soil (Fig. 5).
                              2+
 In the adsorbing system the Ca   ions in the leachate were nearly independent
 of the sulfate ion concentration in the rainfall, i.e. independent of sulfuric
 acid induced rainfall acidity.  The reason for this independence is that the
 solution concentration was largely controlled by the sulfate adsorption pro-
 perties rather than rainfall.
                                                       2-
    This independence was temporary, when the input SO,  concentration was in-
                         2-
 creased, the adsorbed SO,   tended to increase until eventually a new equili-
 brium was established with a higher solution concentration, and base removal
 then proceeded at the higher rate.
    Conversely, when the solution rainfall input was again lowered, base re-
moval preceeded at a higher rate while sulfate desorption occured.  When the
                                                                          2
 evaporation experiment shown in Fig. 7 was conducted using a simulated SO,
adsorbing soil, the time required to reach a new equilibrium was increased by
several -fold.  The actual time, of course, was a function of the specific
K    and k:  of the soil.  These effects point out the possibility that with
 max      '2
sulfate adsorbing soils, increased base removal  may not occur in detectable
amounts until stressed with acid rain for long periods, but if the stress is
                                                           2_
removed, base removal would continue until the adsorbed SO.   decreased to
equilibrium with the new input level.  It should be recognized that these sim-
ulations were based on the assumption of complete reversibility of sulfate
adsorption.  If a portion of the adsorption sites were irreversible the leach-
ing of basic cations after sulfuric acid stress  was removed would be less than
                                    26

-------
                  3.0 .
                0>

                0)
                £4.0


                o
                o
                o>
                  5.0
               ' o
               ,O

                o>
                O

                i
6.0  .
        pH-l/2Pco
          4.0
          3 O
                            20-
          I -5
                          5.0
               45
                                           4.O
                                                   3.5
3.0
                                   F   2-    .1
                              -log  I SO, + CM  in
                               Rainfall, eq / littr
Figure 8.   Model  prediction  of concentrations  of Ca
                    2-
                                                      2+
                                       in leachate
                                           2-
            of a S04    adsorbing soil as a function of [SO.   +  Cl~]

            and pH - %pCa.   Leachate volume is  equal to rainfall.
                                  27

-------
 the  equivalent amount of sulfate adsorbed.
 LONG-TERM  SIMULATIONS
      Several  longer-term simulations were run with the model.   In contrast to
 the  results  reported above, in these cases base saturation and  pH - %pCa were
                                           2+
 .allowed  to respond to gains or losses in Ca   . Results are reported for two
 runs,  one  assuming sulfate adsorption properties similar to those of the As-
 toria  soil (Chao et_ aj_., 1962), and the other assuming very low sulfate ad-
 sorption.  The initial  base saturation was low (20% of permanent change), and
 the  relationship of pH  - JgpCa to base saturation was that shown by the lower
 line (montmorillonitic  soils) of Fig. 2.  Initial conditions and soil proper-
 ties used  in  these simulations are shown in Table 3.
      The rainfall used  was the pH 4.0 rain shown in Table 1.  A total of 1100
 mm rain  per year was simulated, one-twelfth of which was applied each month.
 The  pattern within months included 13 days per month in which rain occured in
 amounts  varying from 1.1 to 22 mm/day (Table 4).   Potential evapotranspiration
 rates  were changed each month and are typical of those in the northeastern
 United States.  Potential evaporation exceeded rainfall for the May - Sept.
 period.
     The predicted solution pH values of a 1:1 soil-water system for a five-
year period are shown in Fig.  9.   The seasonal fluctuations predicted are of
 the  order  of  .25 pH units for the non-adsorbing soil.  This results from changes
 in solute  concentration due to the lack of leaching during the summer period
when evapotranspiration exceeds rainfall.   The increased solute concentration
                                                                       o
decreases  pH due to the "salt effect" discussed previously.  In the SO,   ad-
sorbing soil  these fluctuations are much smaller.  This damping effect is due
                      2_
to the control the SO.    adsorption exerts on solution anion concentration.
     Measurements of pH in soil  water suspensions or leachates often vary mark-
edly with time (Russel,  1961  p 105; Cole and Ballard, 1970).  Smooth seasonal
fluctuations  such as  are observed in the model output are rare, but actual
rainfall  and  leaching  patterns are also highly variable between seasons.  The
effects demonstrated  by the model  are undoubtedly a major factor in inducing
fluctuations.   Effects  of this magnitude are probably realistic and could con-
stitute a major source  of error in experiments attempting to measure field  pH
changes due to acid  rainfall  inputs.
                                     28

-------
Table 3.  Initial conditions and soil properties for a 10-year acid rain-
          fall simulation
                                   2_
                                SO.   Adsorbing          Non-adsorbing
                                    Soil                     Soil
Cation Exchange Capacity
(eq/kg)
Initial Base Saturation
(«)
Depth (meters)
0.10
20.00
0.30
0.10
20.00
0.30
Field Capacity
    (% volume)                     30.00                    30.00
Permanent Wilting Point;
    (% volume)                     12.50                    12.50
Bulk Density
    (kg/liter)                      1.25                     1.25
C02 Partial Pressure (atm)        3 x 10~4                 3 x 10~4
   2_
SO,   Adsorption
    K    (moles/kg)             3.6 x 10"3               3.6 x 10"5
     rriuX
    \ (moles/liter)            1.8 x 10"4               1.8 x 10"3
    Initial S (moles/kg)       8.56 x 10"5              1.42 x 10"5
K   - A1(OHK                    10-33.8                 lo'33'8
 5 p         -3
                                    29

-------
Table 4.  Rainfall  and evapotranspiration patterns used for 10-year simula-
          tions
Month

Jan.
Feb.
March
April
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.

Potential
Evapotranspiration
mm/ day
.033
.067
.533
1.533
3.067
4.367
5.133
4.533
3.323
1.767
.633
.100

mm/month
1.0
2.0
16.0
46.0
92.0
131.0
154.0
136.0
100.0
53.0
19.0
3.0

Rainfall
mm/month
91.7
91.7
91.7
91.7
91.7
91.7
91.7
91.7
91.7
91.7
91.7
91.7
1100.4
                                    30

-------
         58r

      «-  5.6
      ^
         5.01—-
                	Non-Adsorbing So
                        , 2-
                     SO4 Adsorbing Soil
             MAY  OCT  MAY   OCT   MAY  OCT  MAY  OCT  MAY   OCT   MAY
                                   Time
Figure 9.  Simulated time-based fluctuations in soil pH  using pH
           4.0 rainfal1.
                                31

-------
    The total (exchangeable plus solution)  Ca2+  in  the  system  dropped  in  a
stepwise pattern (Fig. 10).  A small gain occured during  the summer  due  to
rainfall inputs with no leaching.  This was followed by a sharp drop in  the
fall.  In this case the loss was lower for the sulfate adsorbing soil  than  for
the non-adsorbing soil, but in the adsorbing soil the loss increased slightly
each year.  This increase  is more apparent in Figure 11 where only the April
30 values are shown.  By the end of the  10-year  period the loss rate approach-
ed that of the non-adsorbing soil.  This occured because  sulfate was adsorbed
from the rainfall, resulting in  higher equilibrium  solution concentrations.
While these effects may be of considerable importance, the graphs must be
interpreted with caution as the  base  loss  pattern  of sulfate  adsorbing soils
depends on the initial sulfate  level.  It  appears,  however, that adsorption
dampens the seasonal  oscillations.
    An interesting effect  is observed  in the  plot  of solution  pH values on
                                         2+                      2-
April 30 (Fig. 12).   Even  though  less  Ca   was  lost from  the SO.   adsorbing
soil than the non-adsorbing soil, the  pH of the  soil-water suspension dropped
at a slightly faster  rate.  The  reason for this  is  not clear.
                             2+
    The net change in soil Ca    per equivalent of excess  acidity in the rain-
fall is about 98% at  the start and  96% at  the end of the  ten-year period for
the non-adsorbing soil.  Base saturation decreased  from 20% to  17.12%.  Fur-
                                                               2+
ther declines in base saturation would rapidly decrease the Ca   loss per unit
of H  added. More precise  information on the nature of the functional  relation-
ship between base saturation and pH - %pCa is required to simulate more accur-
ately the effects in this  region.
                                    32

-------
             3.8
            CO
            QJ

            O
              3.7
           CM
            OJ
            U
              3.6
            o
              3.5
— Non-Adsorbing Soil
      2-
— SO4 Adsorbing Soil

                      OCT  MAY
                                OCT
                                     MAY   OCT

                                      Time


                                        2+
                                               MAY
                                                    OCT
                                                        MAY
                                                             OCT
Figure 10.  Simulated change in total Ca    (soluble  plus  exchange-

            able) in 0.3 meters of soil during a  5-year period

            using pH 4.0 rainfall.
                               33

-------
                 3.8
              CM   3.6
               E
               l/l
               

               "o
              CM
               0)
              u

              To
              +->
               o
                  3.4
      3.2
                  3.0
                            S2-
           • -SO4 Adsorbing Soil

           o- Non-Adsorbing Soil
                      0
                          4        6

                          Years
8
10
Figure 11.
                              2+
Simulated total Ca   found on April 30  in  0.3  meters

of soil stresses with pH 4.0 rainfall over a  10-year

period.
                                 34

-------
             
-------
                                  SECTION 6
                                  DISCUSSION

 LIMITATIONS
     The  major  limitation  of  the model is the exclusion of all basic  cations
 except calcium.   The  system  is complex and it appeared desireable to success-
 fully complete and  test the  model in the present form before attempting to
 include  additional  ions.  The time and resources alloted to the initial pro-
 gram were  fully  utilized  in  completing the version reported here.  Fortunately
 the model  has been  useful in providing a basic insight into the effect of soil
 properties on  leaching of cations due to acid rainfall.
     Inclusion of  additional  cations should not present any serious difficul-
 ties.  They  can probably  bast be handled by the use of the well known Gapon
 type relationship as  shown in eq. (13) - (15) for two component systems.
                                   = C
                               CaX   Cl
                                   = C
                                   = c
                                             (Na+)
                                  .
                               PaY    ^    9+     94-   'a
                               CaX    3 (Ca2+ + Mg2+)
Where [MgX]/[CaX], [KX]/[CaX], and [NaX]/[CaX] denote the ratios of these ions
on the exchange complex, the (  )'s denote solution activities in moles per li
ter, and C,, C?, and C~ are constants for which approximations are available
in the literature.  These relationships plus the concept that the total ex-
changeable plus soluble amounts of these ions are always known in the model,
and the equivalence of soluble anions and cations allow the inclusion of any
                                    36

-------
or all of these ions.  The exchange relationships are the same as those used
by Dutt et_ 
-------
                                            2+
would expect to have a leachate with log [Ca  ]  of -4.52 with no evaporation
(fig. 5).  Actually the K  comprises a significant portion of the cation in
the leachate so that log [Ca2+ + K+]  = -4.52 or [Ca2+]  = 3.16 x 10"5 - K.  For
this semiquantitative calculation we can assume activities equal to concen-
trations and substitute into (16) so that:

                 0.025 - 0.008 + 6(K+)/(3.16 x 10"5 - (K+))       (17)
Solving  (17) we find [K+]  = 9.4 x 10"6 and  [Ca2+]  = 2.2 x 10"5.
     The ratio of dissolved K/Ca is 0.42, or more  than  16 fold the ratio of the
exchangeable ions.  Surprisingly if this same calculation is carried through
for the  pH 3.0 rainfall (Table 2), where log K+ +  Ca2+  from Fig. 5 is -3.0,
the predicted ratio of dissolved K/Ca is 0.067 or  only  about 2% times the
ratio of the exchangeable ions.  Thus K appears to be selectively leached in
this system, but the selective removal of K is diminished by the higher ion
concentration in the more acid rainfall. Similarly the selectivity of K leach-
ing would be depressed by the concentration effect of evapotranspiration.
     As  the ratio of exchangeable K/Ca decreases,  the selectivity of K leach-
ing also decreases.  For rainfall pH 4.0 from Table 2 and assuming no concen-
tration  by evapotranspiration we find that  at an exchangeable K/Ca ratio of
0.0087,  K/Ca ratios would be equal for the  solution and exchangeable compo-
nents.   If exchangeable K is depleted below this level, Ca would be selective-
ly leached.  It should, however, be noted that equation (16) predicts a zero
level of soluble K  at an exchangeable K/Ca ratio  of 0.008.  This apparent
intercept is an artifact due to sites that  exhibit specificity for K  and at
very low exchangeable K/Ca levels (16) becomes invalid (Beckett et_ al_., 1966).
                                                 +                    ?+
However, the concept that selective leaching of K  in preference to €a
ceases and even reverses at very low exchangeable  K/Ca ratios remains valid.
     The assumption of uniform equilibration throughout the depth of soil con-
sidered also imposes limitations.  The results of the model in the form used
for this report should have general validity but will suffer in applicability
to specific situations where various profile depths have widely varying pro-
perties.  Depth effects can be included by  dividing the soil into discrete
layers for which successive equilibrations  are calculated in the model.
Initial  conditions must then be supplied for each  depth.  Precision may also
                                     38

-------
be added by coupling a soil water model to the exchange equilibrium model in
the manner used by Dutt et_ aj_. (1972) rather than simply assuming leaching of
the soil solution in excess of field capacity.
    The relationship of base saturation to Lime Potential utilized here (Clark
and Hill, 1964) is highly empirical.  Similar curves can be derived on a much
more theoretical basis.  Turner and Clark (1964) show that the relationship
can be described by (18).
                                     K                         3
         PH - h p(Ca + Mg) = 1/6 log -^L + 1/6 log [CaX %M9X]       (18)
                                     KjK             [A1X]   CEC

Where [CaX] , [MgX] and  [A1X] refer to the exchangeable ions in moles and CEC
is the cation exchange capacity in equivalents taken as the sum of the ex-
                                                                  -14      o
changeable Ca, Mg, and Al ; the K is the ion product of water of 10    at 25  C,
                                                                  -1 5
K1 is an ion exchange constant for this system determined to be 10  ' , and
                                    3
KAI is the solution product (A1)(OH) .  If Mg is not included:

                                CaX =   (CEC)
Where V is the base saturation.
    Eq. (18) can then be written as

                                   K                         3
               pH - %pCa = 1/6 log  A1RC- -r + 1/6 log 9/8   (V ^ ~    (19)
                                   10"bb-b               (1 - V)
    Thus eq. (19) shows a theoretical functional relationship between base sat
uration and pH - %pCa, dependent only on the value of K. -, .   The functional re-
lationships assumed from Fig. 2 and used in the model as eq. (5) could be re-
placed directly by eq. (19).  A more straightforward approach would be to drop
MgX and use eq. (18) directly along with (20) which simply states that the eq-
uivalents of CaX and A1X add up to the cation exchange capacity.
                              CEC = 3A1X + 2CaX                      (20)
    The eight equations (1), (2), (6), (7), (8), (9), (18), and (20) now form
a set in which, (considering activities to be equal to concentrations), the
                                     39

-------
 unknowns  are; S04(sorbed)>  tS04],  [H+],  [Ca2+],  [A13+],  [HC03"l,  CaX  and  ATX.
 The  use of  the  Lime  Potential  K, as a  constant or as  an  empirical  function  of
 the  base  saturation  is actually  unecessary and the validity of  the system is
 not  actually dependent on the  "constancy" of  the Lime Potential.
     Unfortunately,  I did not recognize the potential  of  this  approach until its
 inclusion would have required  a  major  restructuring of the equilibration  sec-
 tion of the model  so it has not  been included in the  version  reported here.
                            3+
     Controlling solution Al    in the model by means of an "apparent solubility
 product"  of A1(OH)0  is a vast  oversimplification of this equilibrium  involving
                                                                3+    - 3
 Al.   Apparently the  two lines  in Fig.  2 arise from different  (Al   )(OH  )  ion
 products  with the  values for the montmorillonitic soils  generally  ranging
            or r       o/i -j
 between io~°°'° and  10      while most of the non-montmorillonitic soils  are
                   -34 1      -3? R
 in  the range of 10      to  10      .  This difference  rather than  any  direct
 effect of the clay  type appears  to account for the different  lines on Fig.  2
 (Turner and Clark,  1964).   I assume that the  lower equilibrium  levels  of  Al
 in  the montmorillonitic soils  is related to the fact  that this  clay mineral is
 generally found in soils that  are not highly weathered.  For  broad scale  appli-
 cation of the model  it would seem appropriate to investigate  whether  this ion
 product can reasonably be predicted by any of the common soil classification
 schemes.
     The model does not take into account the possible  release of  cations  from
 soil  minerals to the exchange  complex as a result of weathering processes.
 In the present  form it is not  feasible to include this process, but one can
 imagine a more  sophisticated system including equilibria for  various  Al,  Fe,
 and  Si species  that would predict rate of weathering  and cation release from
 soils of  known mineral  composition.
 USEFULNESS
     Even   though several  deficiencies of the model are  recognized,  it  is never-
 theless extremely useful.   The predictions relating to the buffering  or dam-
                         2_
 pening effect of soil SO.    adsorbing properties are  important.  Acidity  ef-
 fects such as pH changes  or cation leaching could lag well behind  the  acid
 inputs of the rainfall  due to  this mechanism, but increased base  removal
would continue after the  acidity inputs ceased.

                                    40

-------
    It is often assumed that soils with low pH and low basic cation contents
would be highly susceptible to atmospheric chemical influence (Jonsson and
Sundberg 1972).  The fallacy of this viewpoint has been previously pointed out
(Wiklander 1973; Wiklander 1974; Malmer and Nilsson 1972).  The output of this
model confirms that soils well supplied with bases are most susceptible to base
loss and that as base saturation and lime potential fall  to low levels acid
precipitation causes leaching of K+ and Al + ions rather than bases. Unfor-
tunately this transition takes place at very low base saturation levels, as
base removal is almost equivalent to rainfall acidity inputs at 20% base sat-
uration or above.
    The obvious method of evaluating rainfall acidity effects in the field is
to measure pH changes.  Several of the problems with this approach are appar-
ent from the output of this model.  For instance the imposition of acid rain-
fall may immediately lower the measured pH value simply due to higher anion
content, even though a similar effect would occur due to addition of rainfall
containing a neutral salt, or as a result of concentration of salts found in
rainfall by evapotranspiration effects.  The seasonal effects on pH shown in
the model are striking.  Unfortunately, in nature these are not smooth annual
cycles but are highly irregular.  Cole and Ballard (1968) show a variation of
0.3 pH units in drainage water through the forest flow over a period of about
4 hours, the pH decreased as conductivity increased as would be expected
from the present model.  The acidic characteristics are further complicated by
biotic uptake and release of both anions and cations.  These processes affect
the acid base status of the soils (Reuss 1977).  Changes in more fundamental
soil properties such as pH - JgpCa or pH - l/3pAl are much more likely to give
meaningful measures of the effect of external acidity imposed on soils than
are pH measurements.
    The initial output from the model indicates that loss of bases from the
system in response to rainfall acidity could be significant when considered
on a time scale of several decades.  The system simulated would have lost a-
                 2+
bout 0.5 moles Ca   per square meter over a decade when subjected to pH 4.0
rainfall at a rate of 1100 millimeters per year.  While subject to wide annual
variation the downward trend in solution pH of a 1:1 soil water system was
proceeding at a rate in excess of 0.1 units per decade for a soil with the

                                      41

-------
properties chosen for simulation.
    One method of shortening the time necessary for field investigations of
the effect of rainfall  acidity is  to apply highly acid artificial  rainfall.
The model output indicates a reasonable prediction of base removal  due to ion
exchange effects, at least to the  extent of inputs at pH 3.0 which  was the lo-
wer limit of the test run.  This can be discerned from the fact that the lines
on Fig. 6 are nearly horizontal  in the region where rainfall pH values are be-
low about 4.0.  Caution should be  observed in interpreting pH values from such
investigations, as highly acid artificial  rainfall  would increase  the solution
ion concentration and pH measurements on the common water suspensions would be
unrealistically depressed.
    The model is strictly abiotic  and does not take into account biological
processes of any type.   It could be very logically used, however, as a module
in a more comprehensive ecosystem  nutrient cycling model.
                                   42

-------
                                 REFERENCES

Anon.  1971.  Sweden's case study for the U.N. conference on the human envir-
    onment.  Air pollution across national boundaries.   The impact on the en-
    vironment of sulfur in air and precipitation.  Stockholm, 1971.

Beckett, P.H.T.  1965.  Cation exchange equilibria of calcium and magnesium.
    Soil Science 100:118-123.

Beckett, P.H.T., J.B. Craig, M.H.H.  Nafady, and J.R.  Watson.  1966.   Studies
    in soil potassium V:  The stability of Q/I relationships.  Plant and  Soil
    XXV:435:455.

Chao, T.T., M.E. Harward, and S.C. Fang.  1962  Adsorption and desorption
    phenomena of sulfate ions in soils.  Soil  Sci. Soc.   Proc. 26:234-237.

Clark, J.S. and R.G. Hill.  1960.  The pH-percent base  saturation relation-
    ships of soils.  Soil Sci. Soc.  Amer. Proc. 28:490-492.

Cole, E.W. and T.M. Ballard.  1970.   Mineral  and gas  transfer in a forest
    floor - a phase model approach.   In:  Tree growth and forest soils.  (C.T.
    Youngberg and C.B. Davey)  Oregon State Univ. Press.   Corvallis  Ore.  p.
    347-357.

Dutt, G.R., M.J. Shaffer, and W.J. Moore.  1972.  Computer model  of  dynamic
    bio-physiochemical  processes in soils.  Univ. of Ariz.  Agric.  Expt.  Sta.
    Tech. Bull. No. 196.

Harward, M.E. and H.M. Reisenauer.  1965.  Reactions  and movement of inorganic
    soil sulfer.  Soil Science 101:326-225.

Hasan, S.M., R.L. Fox, and C.C.  Boyd.  1970.   Solubility and availability of
    sorbed sulfate in Hawaiian soils.  Soil Sci. Amer.  Proc. 34:897-901.

Jonsson, B., and R. Sundberg.  1970.  Has the acidification by atmospheric
    pollution caused a growth reduction in Swedish forests?  Institutionen
    for Skogsproduktion (Department of forest yield research) Rapporter och
    Uppsater (Research Notes), Skogshogskolan (Royal  College of Forestry),
    Stockholm 48 p.

Malmer, Nils, and Ingvar Nilsson.  On the effects of  water, soil, and vegeta-
    tion of an increasing supply of atmospheric sulfur  SNV PM 402E.   Depart-
    ment of Plant Ecology.  Univ. of Lund.  Lund, Sweden.
                                   43

-------
Reuss, J.O.   1972.   Chemical  and biological  relationships relevant to the ef-
    fect of acid rainfall  on  the soil-plant system.   Water,  Air,  and Soil
    Pollution,  7:461-478.

Russel,  E.W.   1961.   Soil  condition and plant growth.   9th edition.   Longmans,
    London.

Schofield, R.K., and A.M.  Taylor.   1955.   The measurement of soil  pH.  Soil
    Sci. Soc.  Amer.  Proc.  19;186-191.

Turner,  R.C.,  and J.S.  Clark.   1964.   Lime potential  and degree of base satu-
    ration of  soils.  Soil  Sci.   99:194-199.

Turner,  R.C.,  W.E.  Nichol,  and S.E.  Brydon.   1963.  A  study  of the lime poten-
    tial^.  Soil Sci.   95:186-191.

Wiklander, L.   1973.  The  acidification of soil  by acid  precipitation.   Grund-
    forbattring 26:155-164.

Wiklander, L.   1974.  Leaching of  plant nutrients in  soils.   1.   General  Prin-
    ciples.  Acto Agriculturae Scandinavica  24:349-356.
                                  44

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing}
  REPORT NO.
    EPA 600/3-78-053
                                                           3. RECIPIENT'S ACCESSION NO.
4.TITLE ANDSUBTITLE
         SIMULATION  OF NUTRIENT LOSS FROM  SOILS
         DUE TO RAINFALL ACIDITY
                                                          5. REPORT DATE
                                                             May 1978
                                                          6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)

        John 0. Reuss
                                                           8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
        Environmental  Research Laboratory-Corvallis
        Office of  Research and Development
        U.S. Environmental Protection Agency
        Corvallis,  Oregon 97330
                                                          10. PROGRAM ELEMENT NO.

                                                             1AA602	
                                                          11. CONTRACT/GRANT NO.
 12. SPONSORING AGENCY NAME AND ADDRESS

        same
                                                          13. TYPE OF REPORT AND PERIOD COVERED

                                                              inhousp
                                                            14. SPONSORING AGENCY CODE

                                                              EPA/600/02
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
 This paper describes  a  simulation model that  provides a quantitative  system utilizing
 established relationships  from soil chemistry to  predict the most likely  effect of
 rainfall acidity on  the leaching of cations from  noncalcareous soils.
 The model utilizes  the  relationships between  lime
 saturation described  by Clark and Hill (Soil   Sci.
 and Turner and Clark  (Soil  Sci.  99:194-199, 1964).
                                                    potential  (pH - l/2pCa)  and  base
                                                     Soc.  Amer.  Proc.  28:490-492,  1962)
               -.-....  v-_ . .  --..  --.... . __,  .-.-.,,  the  equil ibrium between C0? partial
pressure and H  and HC03-  in  solution, the apparent solubility product of AL(OH)3, the
equilibrium of cations and  anions in solution,  the  Freundlich  isotherm description of
sulfate adsorption, and mass  balance considerations,  to  predict the distribution  of
ions between the solution  and sorbed or exchangeable  phases.  Ionic composition of leach
ates in response to rainfall  composition can thus  be  computed.   Ions considered  in the
present version are H+, Ca2+, Al3+,SO/i2~  CL",  and
 The model
 leachate and
                         ,S04d~, CL", and HC03

predicts almost exact chemical equivalence  between basic cation removed  in
   strong acid anions entering the system  in  the rainfall if pH -  l/2pCa
                                                                                        th
 above 3.0, at which point  the  base saturation will  generally not exceed
 pH - l/2pCa values leaching  of anions in association  with  hT and Al3  b
 and these cations predominate  when pH   l/2pCa falls  below 2.0.
                                                                                       is
                                                                          20%.
                                                                        becomes
                                                                      At  lower
                                                                      significar
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b. IDENTIFIERS/OPEN ENDED TERMS  C.  COSATI F;ield/GrOUp
     Rainfall
     Water analysis
     Water chemistry
     Soil  science
     Soil  chemistry
     Plant nutrition
     Ecology
                                              Rainfall  chemistry
                                              Precipitation
                                                (meteorology)  chemistry
                                              Acid rainfall
                                              Soil acidification
                                                                04/E
                                                                06/F
 8. DISTRIBUTION STATEMENT
 Release  to Public
                                              19. SECURITY CLASS (This Report,
                                                 unclassified
                                                                        21. NO. OF PAGES

                                                                          56
                                             20. SECURITY CLASS (This page)

                                                unclassified
                                                                         22. PRICE
EPA Form 2220-1 (Rev. 4-77)   PREVIOUS EDITION is OBSOLETE
                                             45
         U. 5. GOVERNHENI PKINIIFrcTDFKcE. 1976-797-234/194 REGION 10

-------