In instances when both DBH and height was used, the combination was needed for accuracy for that tree species or forest type.  If you had large areas to survey for height, during leaf on period, I would consider LIDAR to give ground surface and to address average height of vegetation.  If you have stand age of each species with height, you can develop your own site index curves, and probably relate that also to your or existing biomass data for each species with DBH and/or height. My earliest experience was involved with a site index study in forestry in Michigan during one summer, but I spent my career in hydrology.   At the time I had some involvement years ago, biomass was not considered as much as it is today.  

For some species or conditions, DBH or height may be best, for some both are needed, but that is typically developed by foresters from data and measurements of those species.  

Tree height more time consuming than DBH measures with D tape.  Clinometer more difficult to read, but easier to carry - probably similar accuracy with care.  Height -DBH relationships for individual species best with fast growing stands, but for old trees, height increase declines more than the diameter increase decline.

I forgot to mention, you may be able to estimate stand height with aerial photo pair interpretation.  Unsure if this tedious task has been automated through remote sensing.  I have also used a tool called the Criterion Survey Laser, and was told by an engineer that this type tool is now much more portable and affordable.  In used it to define the boundaries of a pinochle gully mentioned in 2007 paper on erosion from small gully.

Hi dear Agevi,

I think it is better to use both DBH and Height of all trees if this is possible (I mean if your area is not too large). If your area is too large you can use DBH/Height models to estimate either the height or the DBH but using both DBH and Height is more accurate for biomass studies.

Good luck

Hi,

It depend on the kind of the tree . Also are you interested on the  volume or on the biomass ?

If you are interested to the volume. You will need  to measure the height also. Indeed, in order to establish the equation you will need, as dependant variable, your sample trees volumes. Such volume is calculated through subdivising the tree in logs and calculating the volume of each log. That, need to measure, for each tree, diameters at many height levels in addition to the total tree height.

I thank measuring height and diamaeters along the tree is a little tadious if you will be interested to standing trees. In such case you can use instruments that can either mesure height and diameters (for exemple Criterion RD1000, Digitech Compass, etc...).

Other wise for biomass you will need to measure felled trees. Then, measuring the height will be easier.

You will find on the link a good FAO book related to the developpement of biomass equations (in french, I suppose that you can find an english release).

Good luck.

You are best off using both height and diameter if you are working in forests with diverse stand densities. If you are working in monoculture stands with identical stand densities (trees per acre - basal area per acre etch.) and similar ages then you are probably fine developing a height diameter regression relationship and using it to predict height. The reason for this is that trees will preferentially put energy into height growth, so as competition increases (with increasing stand density) we expect the height diameter ratios to increase. Clearly then if you are working in stands with different levels of competition you should expect the height diameter relationships of your trees to be different. This concept will also apply to stands that have distinct canopy classes as the dominant trees will have different height diameter relationships than the intermediate and suppressed trees will. If you develop a height diameter regression model that accounts for the relative competition (for light) experienced by your subject tree then you would be fairly safe in using that to predict heights. 

You are best to use both height and diameter brest height model to determine the biomass  studies while considering a large area of diverse tree stand densities

Allometric equations that just use DBH can be quite precise and unbiased if the you have ensured that they have been tested or calibrated for that one area. You need more variables - height, taper, etc., if you are extrapolating over a larger area. If the area is small, I would probably recommend just using DBH ...but use the time saved by not measuring height to destructively sample a range of trees to calibrate the equations for the local area.

Obviously, DBH and height offer better results.

But using DBH and height is more accurate only if you measure the height, because if you use a computed height depending on DBH, the output is a direct function of DBH, so there is no gain in accuracy.

On the other hand, depending on the species characteristics, you can obtain good results only with the DBH. For some species the model of height as a function of DBH works pretty well, but for other it doesn't.

The best way is to use clinometer or hypsonometer to measure individual heights.

There are some electronic laser hypsonometers (e.g Vertex series from Haglof) that offer a good accuracy, are quite small and very easy to use.

If it takes too much time to measure all the individual heights in the field, you can sample your population and measure the height only for a certain amount of trees and then build your own regression that will feet your species and your study area, with good results. Thus, the model will be calibrated for your need.

Dear Agevi

The issue also revolves around the tree species. In some case diameter is also capable to represent the tree in complete sense, in such case one can go for equation with dbh only. 

The equation containing dbh and height needs to be compared first and also validated for your situation through fresh data. That will decide the correctness of the model, inspite of statistical tested. More to that pl be cautioned that height measurement is not only tedious rather involve more error in measurement. So if you are getting more valid result with only dbh, you may opt this. 

Thanks

Dear Agevi

You can use allometric equations provided the equations are developed by considering the similar situations where it is being used. Before using the equations you have to consider the following

1. Most of the biomass equations/allometric equations are  species specific and often localized in nature. 

2. Many a times the biomass of the same species differ with spacing, age and site class.

3. Variables used in allometric equations could also important.

IN my opinion it is better for you to measure height and DBH and develop your own allometric equation of the agroforestry situation 

Best wishes

Good day! LIDAR is also an effective detection system that works on the principle of lasers to measure both ground height of vegetation. 

In general, two-variable (D & H) biomass model can give more accurate estimate than one-variable (D) model. But if you are interested in total aboveground biomass only, not involving stem, branch and foliage biomass components, the difference between one- and two-variable models will be very small. In the case of applying two-variable model, it is better to use the observed values of tree height than the estimated values from D-H equation. Of cause, just like Legrand said, if your area is too large you can use D-H model to estimate the tree height.

Hope it is helpful.

I believe that in practice, the inclusion of height as independent variable may produce greater estimation errors due to lack of accuracy in its estimation even using inclinometer or hypsometer. As the measurement of DAP is very accurate, its exclusive use in the models seems very appropriate.

The problem here is that you are referring to an agroforest setting, if the papers you are considering were not developed with agroforest data the relationship between height and diameter may differ. For example, dense forest will generally have more height by diameter than sparse agroforest. Check that out first. If this may be a problem, utilize the two variable models. If it will not be a problem (tree species height-diameter not impacted by density or diameter only model included agroforest data) then you should use the diameter only model.

Something very important to take into consideration is the species the equations were developed for, and the species you have in your sites. I also agree that DBH based equations can be more applicable since you can measure DBH with more accuracy than height. The desire level of precision is also important to take into consideration, my guess is that you need a rough estimation, then you can be more flexible about the methods to use. 

I suggest than using standard allometric equations, standard sampling procedures will provide better estimate than using models. allometric equations are species and site specific. agroforestry systems is combination of different species of trees (i.,eg Multistorey, SALT 1-4, etc...) the use of allometric equations therefore might increase your  deviation to the actual values. Standard sampling procedures and the use of Blow-up factor formula will be better.

Direct measurement should always be more precise and accurate than a regression equation estimate.  but direct measurement of height when a tree is over 3 or 4 m in height is time consuming.  so one needs to consider how much time you will spend versus the value of more precision and accuracy.  also, when  you use a regression equation the accuracy and precision decrease as the value of the X variable gets further from the average for the data used to develop the equation.  Of extreme importance is how representative of the trees to be measured were the trees used to develop the equation.  A consideration is to estimate dbh and height of  a sample of the trees to be measured.