Contents 

Estimating Trophic State From Secchi Depth

Using Empirical Chlorophyll-Secchi Disk Relationships

Literature Cited

Return to:
  Secchi Disk Page
  Methods Page
  Dip-In Home Page

The following text has been excerpted and modified from from copyrighted material.  Material in this section should not be reproduced without permission of the North American Lake Management Society (www.NALMS.org).  The reader is encouraged to read the original material for more details on Secchi disk theory and methods.  Material should be cited as:

Carlson, R.E. and J. Simpson.  1996.  A Coordinator’s Guide to Volunteer Lake Monitoring Methods. North American Lake Management Society.  96 pp.

Secchi Depth, Algae, and Interfering Substances

The limnologist and volunteer are usually interested in using the Secchi depth to estimate the amount of algae in the water.  This relationship is based on the idea that algal particles affect the penetration of light into the water and therefore, the Secchi depth.  In essence, the light entering the water will be either absorbed or scattered by particles, dissolved colored matter, and the water itself.  As the attenuation of light by dissolved colored matter or particles increases, the Secchi depth decreases.  This inverse relationship produces the typical hyperbolic curve when Secchi depth is plotted against potential attenuating substances, such as algal chlorophyll, color, turbidity, or suspended solids.

What makes a Secchi disk reading useful depends heavily on the use of the data.  If the intent of the measurement is simply to obtain transparency information, then the Secchi depths can be used without problem.  Preisendorfer (1986) argued that a Secchi disk value should be considered nothing more (and nothing less) than a simple visual index of the clarity of a body of water.  He cites the work of Arone (1985) as an example of a legitimate use of Secchi data.  Arone used 96,000 Secchi depths to produce an atlas of Secchi depths around about 50% of the world's coastlines.  Such a use of the Secchi disk information makes no further assumptions as to the cause of the varying transparencies:  only further studies with more sophisticated instrumentation and methods could reveal the causes of the variations.

Estimating Trophic State From Secchi Depth

If the intent of the volunteer monitoring program is not to use transparency per se or to use it in trend analysis, but to use the Secchi depths as surrogate measures of algal chlorophyll or algal biomass, and subsequently, as an indicator of the trophic state of the lake, then a number of other potential interferences become very important.  

The definition of trophic state may vary, but plant chlorophyll pigments is often assumed to be a major indicator of trophic state.  In theory, algal chlorophyll should be able to be estimated from Secchi depth because it is a substance that attenuates light in the water column.  Chlorophyll is also packaged in cells, and these cells absorb and scatter light.  Secchi depth, therefore, should be able to be used as a surrogate estimator of algal abundance, either by producing empirical relationships between Secchi depth and chlorophyll or by deriving the chlorophyll concentration based on the theoretical relationship between transparency and chlorophyll.

If all variables related to the estimation of Secchi depth and chlorophyll were constant, then a generalized and simplified Secchi depth equation could be represented as

Since such knowledge of the values of these variables is often difficult to obtain independently, usually an empirical relationship between Chl and 1/SD is derived instead by plotting and then regressing the logarithm of Chlorophyll against Secchi depth.  For such an empirical equation to be useful, k_c, k_w, and all the other factors involved in a Secchi disk measurement (glare, sun angle, size of disk, etc.) must be constant within the data set.  With data sets of single lakes over limited time periods, these assumptions may hold true.  However, using these derived relationships over a larger group of lakes, or even over other time periods in the same lake, invites problems.  Variations in the attenuation of both chlorophyll, other attenuating substances, and other factors cause scatter in these relationships.  

Many Secchi disk-chlorophyll relationships exist in the literature, often derived for single lakes or regions .  These regression models vary considerably in both the slope and intercept, emphasizing that the relationship between Secchi disk transparency and trophic state variables such as chlorophyll is highly variable and may vary for a number of reasons.  These varying factors may be related to the method of measurement of Secchi depth or chlorophyll, to variation in the amount of other attenuating substances such as non-algal turbidity or dissolved colored substances such as humic acids, or to the nature of the algae themselves such as the size or species of the algae or the amount of chlorophyll packaged in the algal cells.  

Using Empirical Chlorophyll-Secchi Disk Relationships

Sometimes coordinators cannot avoid using empirical relationships to predict chlorophyll from Secchi depth values.  Programs are often charged with determining the "trophic state" of lakes, but have only Secchi depth values available.  In these cases, the coordinator has no choice by use empirical relationships between chlorophyll and Secchi depth.  However, it is the responsibility of the coordinator to emphasize and evaluate the potential error involved in such a use of Secchi depth.  The coordinator should use these empirical models with (a) extreme caution and (b) with knowledge of their limitations.

Empirical chlorophyll-Secchi Disk relationships work best in situations where chlorophyll is the dominant attenuating substance.  In these cases, a regression of 1/SD against chlorophyll may yield a useful predictive model.  Be sure to plot the data as well as derive a predictive equation.  In many instances, this relationship can have considerable heteroscedastcity, that is, that the variance increases with increases in chlorophyll.  In order to minimize this effect, often the data are transformed into a logarithmic form.  Probably most extant models use log-transformed data, often using log Secchi depth versus log chlorophyll.   A plot of the data should produce a straight line.  If the line is not, then some variable, probably other than Secchi depth or chlorophyll is affecting the model.

At this point, you could try using residual analysis to detect the interfering variables.  Residual analysis involves plotting the deviations of the actual data points from the predicted value against various other measured variables.  These plots can reveal if there are any other variables that should be considered and perhaps isolated.  Once confounding variables have been identified, they can be isolated by lumping the Secchi depth data into categories, such as natural lake and reservoirs or into geographic regions, such as ecoregions (Heiskary, 1989).  Alternatively, these ancillary variables can be incorporated into the predictive equation itself using multiple regression techniques.  Equations do exist that can be used to factor out the interference of water color.  In the case of volunteer monitoring programs, the use of geographic or lake type categories might be most useful because they do not require the simultaneous measurement of ancillary variables.

One simple method for isolating highly turbid lakes is to use the color of the water, as measured by the volunteer, to distinguish between lakes having high non-algal turbidity.  The Ohio NEFCO and CLAM programs use a color strip to estimate the color of the water.  There is evidence from a number of northern Ohio lakes in the Ohio-NEFCO program that the relationship between chlorophyll and the chlorophyll predicted by Secchi depth deviates sharply when the reported water color is brown.  These brown-colored lakes are considered to be turbidity-dominated, and Secchi depth data from these lakes could be reported separately from those lakes in which algae are thought to dominate light attenuation.

It is very important that the coordinator recognize and deal with the problems of using Secchi depth as an estimator of chlorophyll concentration.  There are simply too many variables that must be either known or assumed to be constant to allow any prediction with any certainty.  Of course, at times, in some lakes, chlorophyll can be predicted with accuracy using empirically-derived relationships between chlorophyll and Secchi depth, but it is worth remembering that even a stopped clock is accurate twice a day.  The only method to assure the accuracy of the prediction is to continually recalibrate the relationship by measuring chlorophyll simultaneously with Secchi depth.  Of course, if this is done, there is little necessity to measure transparency as well.  

This discussion is not meant to disparage the use of Secchi disks, only to emphasize that the disk measures transparency of the water, and only when chlorophyll is the sole attenuator of light (or k_w known) under optimum (and constant) viewing conditions can Secchi depth predict chlorophyll concentrations consistently and accurately.  Since these conditions are rarely met, it is better to use the Secchi disk for the purpose for which it was designed; a cheap and simple measure of water clarity. Clarity is not a trivial measurement, especially when it can be compared with other values.  It can be used for trend analysis in a waterbody, for location of differences in transparency within a waterbody, and to compare waterbodies.  In addition, when coupled with the measurement of other variables such as chlorophyll and total phosphorus, it can provide meaningful information on the relationships between these important trophic state variables.

Summary

The Secchi disk is and should be the cornerstone of volunteer lake monitoring programs.  It is inexpensive and provides useful data.  However, it does have a number of technical problems that need to be addressed.  Many of the problems can be minimized by standardizing the equipment and carefully training the volunteers.  

Problems of interpretation generally arise when the transparency data is used for purposes for which the Secchi disk was never intended.  The Secchi disk measurement is subject to numerous interferences related to non-algal or non-chlorophyll materials in the water.  Although empirical relationships can be established in some lakes and regions relating Secchi depth to algal chlorophyll, these relationships can change seasonally and between lakes.  Coordinators should use these relationships with caution and re-calibrate the relationships often.

Literature Cited

Arone, R.A.  1985.  Coastal Secchi Depth Atlas.  NORDA Rpt. 83.  Naval Ocean Research and Development Activity.  Ocean Science Directorate.  NSTL, Mississippi  39529-5004.

.Canfield, D.E. Jr. and L.M. Hodgson.  1983.  Prediction of Secchi disc depths in Florida Lakes:  impact of algal biomass and organic color.  Hydrobiologia.  99: 51-60.

Heiskary, S.A.  1989.  Lake assessment program:  a cooperative lake study program.  Lake and Reservoir Management.  5: 85-94.

Preisendorfer, R.W. 1986.  Eyeball optic of natural waters:  Secchi disk science.  NOAA Tech. Memo. ERL PMEL  67.  90 p. NTIS PB86 224060/AS.  

An illustration of heteroscedastcity as the variance of the data increases as the inverse of Secchi depth increases.  Data from Canfield and Hodgson (1983).

Taking the logarithm reduces, but does not eliminate heteroscedastcity in the data set above.  However, it does appear that there is an upper "ceiling" from which data deviate.