tailieunhanh - Statistics for Environmental Science and Management - Chapter 6

Trước-sau-kiểm soát tác động (BACI) thiết kế lấy mẫu thường được sử dụng để đánh giá những ảnh hưởng của một sự thay đổi môi trường được thực hiện tại một điểm được biết đến trong thời gian, và được gọi là 'nghiên cứu tác động tối ưu thiết kế của Green (1979). Ý tưởng cơ bản là một hoặc nhiều trang web có khả năng ảnh hưởng được lấy mẫu trước và sau khi thời gian của tác động, và một hoặc nhiều các trang web kiểm soát mà không thể nhận được bất kỳ tác động được lấy. | CHAPTER 6 Impact Assessment Introduction The before-after-control-impact BACI sampling design is often used for assessing the effects of an environmental change made at a known point in time and was called the optimal impact study design by Green 1979 . The basic idea is that one or more potentially impacted sites are sampled both before and after the time of the impact and one or more control sites that cannot receive any impact are sampled at the same time. The assumption is that any naturally occurring changes will be about the same at the two types of sites so that any extreme changes at the potentially impacted sites can be attributed to the impact. An example of this type of study is given in Example where the chlorophyll concentration and other variables were measured on two lakes on a number of occasions from June 1984 to August 1986 with one of the lakes receiving a large experimental manipulation in the piscivore and planktivore composition in May 1985. Figure illustrates a situation where there are three observation times before the impact and four observation times after the impact. Evidence for an impact is provided by a statistically significant change in the difference between the control and impact sites before and after the impact time. On the other hand if the time plots for the two types of sites remain approximately parallel then there is no evidence that the impact had an effect. Confidence in the existence of a lasting effect is also gained if the time plots are approximately parallel before the impact time and then approximately parallel after the impact time but with the difference between them either increased or decreased. It is possible of course for an impact to have an effect that increases or decreases with time. Figure illustrates the latter situation where the impacted site apparently returns to its usual state by about two time periods after the impact. As emphasised by Underwood 1994 it is desirable to have more .