2. Interpretation of results: The obvious message from the results section and the tables and graphs is that all three sampling techniques produced similar results. One method, the Line Transect, encountered all ten of the different kinds "species" of ground cover at the site. Consequently, Line Transects may be more sensitive to sampling less prevalent landscape features. Strip Transect and Plot Survey both failed to yield any data for three of the parameters. This fact raises questions as to the adequacy of Strip Transects and Plot Surveys conducted with the degree of sampling contained herein. Evaluation of the number of samples required to produce a statistically valid sample size lies outside the scope of this investigation.
In spite of the potential limitations of Strip Transect and Plot Survey methods, all three methods were in agreement relative to the three major ground covers on the survey site. Natural litter was most prevalent, grass was second and bare soil the third most dominant ground cover. Even the percentages were in relatively close agreement for these parameters. Both the Strip Transect and Plot Survey found forbs to be the fourth most dominant ground cover. The Line Transect method placed forbs as fifth, with trees in the fourth spot. Trees were the fifth most encountered ground cover using the Plot Survey method, but were totally missed on the Strip Transects. Since both Line and Strip Transects were performed over the same one hundred meter plots, this could point out a shortcoming of the Strip Transect methodology. Shrubs appeared to be equally identified by all techniques, falling in a tie for sixth using Line Transects and Plot Surveys and seventh employing the Strip Transect method.
All three methods found human litter and cactus to be among the least (or absent) ground cover. Only rock (which ranged from a high of fifth most prevalent on the Strip Transect to eight on the Line Transect) and animal litter (which was sixth on the Strip Transect and not encountered on the Plot Survey) appeared to differ in percent coverage between the three methods.
3. Contextualization of the Results: Ecological scientific inquiry often involves a comparison of the similarities or differences between sites or populations within sites. One method is generally used to sample two populations to determine if they are actually different. This investigation involved sampling the same habitat (population) to determine if three different methods would produce the same or similar results.
In order to statistically evaluate the results, each method was treated as a different population. The null hypothesis is: There is no statistical difference detected between the three methods employeed in measuring the frequency of occurrence in the population. With our limited class time constraints, it was not possible to procure enough survey data to negate our null hypothesis. Two degrees of freedom for each method produced results that were not sufficiently rigorous in this regard. By subdividing our one hundred meter transects into two transects of fifty meter length, we could perform Students' T test in a serial fashion to compare each of the three pairs of possible combinations. There is still inadequate data to negate the null hypothesis. Therefore, no statistically significant difference between any of our three "populations" was shown.
A visual comparison of the tables and graphs in the results section, and a glance at the mean averages of frequency for our population occurrence confirms these calculations, and is probably much more meaningful for the viewer.
In applying the results of our study, the following conclusions and recommendations are made. There is no particular order to the comments, but an attempt has been made to be comprehensive. All three methods require a comparable amount of time to perform, but the Plot Survey method requires an additional amount of time to set up and take down. The Plot Survey and Strip Transect methods may be more suited to sampling in a uniform habitat, or one where there is an even distribution in order to avoid under sampling of minority constituents. The Line Transect and Strip Transect methods require the least equipment. One hundred meter tapes are very expensive, but if not available, any suitable string or rope that can be marked (possibly in advance) into one meter sections will suffice.
All three methods are fast paced and interesting. Each method should hold the attention of student researchers. Comparison of two populations is preferable and makes statistical analysis much easier. Instruction in monitoring techniques prior to arrival in the field will eliminate on site confusion and produce more uniform survey results. The Plot Survey method causes the most site disturbance. If the same plots are monitored repeatedly, site disturbance could play a significant role in distorting out year data. The Plot Survey method "looks the most scientific".
4. Analysis of Errors: Errors incurred in our project are of two types. First, we should have kept our analysis to only two methods and performed more replicates of each. In that way we might have gained sufficient data to negate our null hypothesis and show a statistically rigorous difference in results obtained with the two methods. Additionally, statistical analysis of our data would have been much more straightforward in examining two populations rather than three. Second, we failed to calibrate our reading of the survey point prior to data collection. This likely produced discrepancy in data collection and may have negatively affected the project. Inherent in any data collection is a certain amount of subjectivity. This is especially true when data is collected by more than one researcher. Practicing before arrival in the field could have enhanced uniformity and objectivity of data collection. Finally, we probably made mistakes of which we are not aware.
Feedback and suggestions for enhancement of this study are sought and encouraged from all readers and reviewers.
5. Summary of Conclusions: By a simple visual comparison and examination of the mean averages for the three methods, our investigation shows that any of the three methods would obtain comparable results. Therefore, the method utilized in determination of frequency of ground cover in an evergreen forest habitat can be based on the set of criteria and goals that surround the study.
6. Suggestions for Future Research: Since lack of adequate data handicapped our ability to perform rigorous statistical analysis, the obvious next step is to secure more sampling using all three protocols. It would be especially interesting to determine the level of data collection required for Strip Transect and Plot Survey methods to sample less dominant "species". It would also be interesting to determine if the Line Transect method sustained the current distribution, or if the initial two transects were anomalous. Finally, it must be determined if these results are consistent in other habitat types, or only applicable to forest biomes.