Air pollution

Diatom critical load model

The impact of acid deposition on lake ecosystems is clearly shown in diatom diagrams generated using palaeolimnolgical techniques. As such the point in time when acidification began to impact a site is readily distinguishable, and in the majority of cases it began in the 19th century with minimum pH values being seen in the 1980s.

It can be argued that the point at which change is first recorded in the diatom record of a lake is also the time that the critical load for that lake is exceeded. If we assume that acidification is a function of lake sensitivity to acid deposition and that this is constant at all sites, an empirical model of acidification can be generated by comparing the current acidification status of lakes of differing sensitivity and acid deposition.

Acid emission and deposition processes Calcium (Ca++) is used in the model as the measure of sensitivity to acidification. Acid deposition is represented as the mean 1986-88 total sulphur (S) deposition data. Acidification in the diatom record is defined as a change in floristic composition towards a more acidophilous assemblage.

The acidification status of a set of lakes is plotted against Ca and S deposition expressed as acidity (Keq H+ ha-1 yr-1). Acidified sites all have a relatively low Ca to S deposition ratio. A logistic regression model is used to identify the value of the ratio between Ca and S deposition (Ca:S) where there is a 50% probability of a site being acidified. Battarbee et al (1995) showed that a Ca:S ratio of 94:1 was the critical value for a data set of 41 upland UK lakes.

This ratio is the critical ratio and can be used to set the critical load for any site by calculating the acid deposition for a 94:1 ratio. There is uncertainty in the extent to which acidification has altered Ca levels. The diatom critical load model takes the most cautious approach and uses the Henriksen F-ratio to calculate 'pre-industrial' Ca values (Henriksen et al. 1986). The diatom critical load is then calculated as follows:

  1. Use the Henriksen 'F-factor' to calculate pre-industrial Ca++ values (Cao)
  2. Use the critical ratio (CR) to calculate the critical load for S as = Cao / CR
  3. Re-express as Keq S ha-1 yr-1
  4. Calculate exceedance values from the difference between critical load for S and current S deposition

The diatom critical load model depends on the accuracy of the Ca:S ratio. This ratio can be validated by analysing sediment cores from a range of non-afforested sites that have a wide range of Ca:S ratio values. Current validation exercises indicate that the ratio is robust.

Ideally the model would use the Ca value for a site at the point of acidification (Ca[x]), however this is generally not known at most sites. As a precautionary approach the model uses Cao which gives lower critical loads (i.e. less deposition is needed before a site is exceeded or damaged).

The model has only been calibrated using non-afforested lakes, so it is assumed that the critical loads apply equally well for streams. This assumption is un-testable as stream sites do not have diatom stratigraphic records. The model also does not take into account nitrogen (N) deposition.

Critical loads calculated using the diatom model are complimentary to those calculated using the SSWC and FAB models, which are routinely used for critical load mapping in the UK and Europe. SSWC and FAB are constrained by the use of a pre-determined ANC (acid neutralising capacity) value, ANC[crit]. In the UK, ANC[crit] = 0 µeq ANC l-1 = ANC[0], which is determined by the ANC value where there is 50% probability of damage to Brown trout (Salmo trutta). As such the SSWC and FAB models do not calculate the critical load for a site because the ANC at which exceedance occurs varies from lake to lake. The diatom model is not constrained in this way as it is not limited to a pre-determined ANC[crit] value and indicates the critical load threshold beyond which a site starts to be affected by acid deposition.

The advantage of using the SSWC and FAB models is that they can be used to set critical loads for target organisms. ANC[0] is defined for brown trout but any ANC[crit] can be calculated using organism repsonses to ANC. The diatom critical load can not be used in this manner.


  • Battarbee R.W., Allot T.E.H., Juggins S., Kreiser A.M., Curtis, C. and Harriman R. (1995) Critical loads of acidity to surface waters: and empirical diatom-based palaeolimnological model. Ambio 25, 366-369.
  • Henriksen A., Dickson W. and Brakke D.F. (1986) Estimates of critical loads to surface waters. In: Critical loads for sulphur and nitrogen (Ed. J. Nilsson). Nordic Council of Ministers, Copenhagen, pp. 87-120.