Any sample may normally have some type of “representative” quality. However, if your question is about a “forest’s representativeness” the size of the surveyed area is of limited importance. What is more important is the variability of the vegetation within the entire forest, and the degree to which you have managed to capture this variability in your sampling, so that your samples optimally “represent” the entire forest (i.e. many samples spread over the forest in strategically planned manner). 0.1 ha seems a rather small area for sampling, especially in tropical forests. But how did you get to this 0.1 ha number? Have you already done the survey, or are you still planning on doing it?

Dr. Roland Cochard, 

Thank you for sharing your opinion. I am still on the planning stage in designing my sampling method. My study site is a primary forest of an island which is why i choose to do 0.1 ha for plant survey. However, i do need opinion in confirming on this matter. 

There is nothing I could "confirm" here, neither in terms of "yes" nor "no". I would urge you to think carefully about 1.) what questions you want to address exactly with your survey, and 2.) how much time you will have available to do the survey (do you have any experience there already? -> if not, make some pilot trials). These two things will essentially determine your approach, including the total area sampled (-> if you find that you could sample more you may consider doing more to improve the database/representativeness). If you really want to describe the "forest" of the entire island (is this indeed the aim?) you should think about what types of forests there may actually be (using various available information, including literature, satellite images/maps, preliminary field visits), e.g. beach forests, freshwater and brackishwater swamp forests, semi-deciduous forests on hillsides, ->...whatever there may or may not be..., and how you think about stratifying the sampling (and its corresponding time and effort...). Cf. answer by Ms Almaarofi. Have a look at various literature (e.g. http://eu.wiley.com/WileyCDA/WileyTitle/productCd-0471490938.html) to see what other people did or suggest in similar conditions. I hope that helps. best, R

For tree species composition in moderately divers forest, your total sampling area, even for unbiased distributed sub-samples, seems an order of magnitude to small. An example is attached. If you are not studying botanical composition, but let say timber volume, or cover changes, I may provide other real world examples.

Dear Dr. Hein,

Thank you for the example that you have attached.

Thank you Dr. Sama and Dr. Depish for the information. I appreciate it a lot.

It depends on the sampling methodology and total area of the location.

Regards

It depends mainly on the stipulated error you are working with. This error will be compared to the sampling error that depends on the sampled size (0.1 ha in your case). It also depends on the variable studied, the type of vegetation (plantations or natural forests). Natural forests, mainly tropical forests, show greater variability and, consequently, the sampling effort to achieve sample sufficiency will be greater than in plantations. It also depends on the sampling system adopted: sampling procedure (randomized, stratified, in conglomerate, etc.) and sampling methods (plots of variable areas and fixed areas such as square, rectangular, circular, etc.). There are formulas for calculating the sample size for each procedure. All this is detailed in Forest Inventory Books.