Principles of Vegetation Measurement & Assessment
and Ecological Monitoring & Analysis


Veg Sampling
 © 2009 University of Idaho
 All rights reserved.

 Web Design - CTI

Estimating Biomass and Double Sampling

Return to Module Overview

<== Previous Lesson                                                                                       Next Lesson ==>

measuring Biomass

Estimating Biomass

It takes a lot of time to clip many plots to make good estimates of biomass. However, with a little training most field technicians can become skilled in estimating the amount of biomass in a plot.

  • To accurately estimate the amount of phytomass in a plot, the observer must spend time training.
  • The training procedure basically entails weighing representative units of a plant and establishing an “eye” for what 5-, 10-, or 15-grams etc., of forage looks like.
  • Estimating biomass is both visual and tactile.  Good estimates generally require looking at the plant or plot and then "feeling" it to assess density.

These examples are from
Chapter 4 of the
National Range and Pasture Handbook

This procedure of estimating biomass can be easily accomplished for small herbaceous plants. It is more difficult to gain excellence in estimating shrubs and trees but, it is possible. Consider this example below:

The Reference Unit Method is a slight modification of the Direct Estimation Method described above that is particularly well suited for shrubs. In this method one simply clips a unit of the plant and carries it along to each plant instead of learning to recognize a specific unit of weight:

  • A small unit of a plant (such as an average sized branch, see figure to right) is designated as the reference unit and clipped from the plant.
  • A reference unit should be 10-20% of the foliage weight of the average plant.
  • The reference unit is then held up against plants for which phytomass estimates are required. The number of reference units in other individual plants being examined is recorded.
  • The weight of current season’s growth or total mass of the reference unit is then determined.
  • The weight of estimated plants = number of ref. units * wt of the ref. unit.
  • The techniques works well for some shrubs, but is not well suited for compact, dense, unsegmented growth forms.

The accuracy of the observer’s estimate depends on:

  • The experience of the observer. Well trained technicians with a good deal of field experience can estimate the amount of forage in a plot with little error.
  • The alertness of the observer. Accurate estimation requires significant concentration. Accuracy often decreases at the end of the day when observers are tired, hot, or hungry.
  • The vegetation type. Some plant types are simply easier to estimate than others. For example, bunchgrasses are often easier to estimate than sod-forming grasses.

 Double Weight Sampling

Estimates of biomass can be calibrated by clipping a few plants or plots after estimates are made. This procedure is called 'Double Sampling' and basically requires that the field technician estimate the weight of several plots and then clip a few plots to determine the accuracy of estimates. Then, estimated weights can be adjusted to reflect clipped weights.

For example, a comparison of estimated weights and clipped weights might reveal that the observer consistently underestimates the weight of a plot by 75%. Once this is known, the estimated weights can be adjusted to reflect a more realistic and slightly higher amount of biomass per plot.

The advantage of double sampling is that it takes a lot less time to estimate the weight in a plot than it does to clip it. Therefore, many more plots can be examined in a landscape, pasture, or management unit.

How many plots should be clipped? In this technique many plots will be estimated and several will be clipped. The number of plots to be clipped depends primarily on the variation in phytomass from plot to plot and the accuracy of the observer’s estimates. A good rule of thumb is to harvest at least 1 plot for every 7 estimated. Further guidelines include:

  • Clip enough quadrats so that some quadrats represent the least amount of phytomass likely to be encountered on the site and some quadrats represent the greatest amount of phytomass on the site.
  • Each quadrat should be estimated first and then a random procedure (e.g., a coin toss or random generator in a computer) should be used to determine if the plot needs to be clipped. If this is not done, the observer will tend to estimate the plots that need to be clipped more carefully than those that are not going to be clipped.
  • Ideally, the observer should never see the weight of the clipped plot or the observer will adjust the weights of subsequent plots. In double sampling methods, it is more important to be precise and consistent than it is to be accurate. However, the practice of predicting, clipping, then weighing does improve the accuracy of an observers guesses over time.
  • There is disagreement over whether an observer should try to estimate dry weight or fresh weight in a plot. Generally, fresh weight is estimated because it seems more relevant in the field.  But, this requires collecting field samples to weigh and correct estimates to dry weight.  Samples may even need to be collected throughout the day as plants become drier.

Adjusting estimated weights with double sampled plots:

  • A short-cut procedure based on the average difference between the estimated and clipped plots, is given on pages 106-107 of the Interagency Handbook,
    Sampling Vegetation Attributes
  • The preferred procedure is to conduct a regression analysis. All plots that were estimated and then clipped are used to create a regression line: y= a+bx.
       y is the clipped weight of a plot
       x is the estimated weight of a plot
       a is the y-axis intercept (a constant that is added to each estimate)
       b is the slope of the regression line

    Computer calculation. Most computer spreadsheet programs can develop a regression line.

    Hand calculation. The procedure for calculating a regression line can be found in most basic statistics text books.

    For information on how to calculate and use regression models visit:
  • After a regression line is created, all of the estimated weights can be adjusted to improve the accuracy of the guesses.
    Once again, y= a+bx   with “y” being the adjusted weights and “x” the estimated weights.

    For example:
    If your calculated regression line is: y = -20 + 1.5x
    And, you estimated value (x) is 30 g/plot in the field
    Then y = -20 + (1.5*30) or your adjusted estimate is 25 g/plot
  • Express weights on a dry weight basis. The clipped weights and the adjusted weights should then be put on a dry weight basis with the procedure discussed in lesson 7_3 under the 'clip-and-weigh' or 'harvest' method.
  • Convert weights to meaningful units. When all weights are adjusted and averaged the result will be phytomass in a small area. Guidelines for converting plot weight to lbs/ac or kg/ha are located in lesson 7_3.

Summary Questions

  1. What factors reduce the potential accuracy and consistency of field biomass estimates?

  2. What is the advantage of using double sampling methods over clip-and-weigh methods?

  3. In the following example, I estimated the weight of 8 sagebrush shrubs. Then, clipped all the current season's growth form each plant to get a actual weight:

1 2 3 4 5 6 7 8
Estimate (g) 20 45 150 600 30 45 90 10
Actual (g) 15 50 165 553 28 70 110 25


How accurate was my estimate? How far off from the actual weight, on average?

Next, I calculated a regression equation between my clipped and estimated pots:

Note that the R2 of my estimate is .98 or 98%.  This is very good. If I had done a perfect job of estimating my R2  would be 1.0 or 100%.  The closer to 1.0. the better.

● If I went back to the field and estimated the weight of four more plants as:

1 2 3 4
Estimate (g) 24 350 63 82


What would the adjusted estimates for the 4 plants be if I applied the regression equation from my graph to these estimates?

**Click here for answer

Return to Module Overview

<== Previous Lesson                                                                   Next Lesson ==>