29.        Salvage Logging. When wildfires ravage forests, the timber industry argues that logging the burned trees enhances forest recovery. The 2002 Biscuit Fire in southwest Oregon provided a test case. Researchers selected 16 fire-affected plots in 2004—before any logging was done—and counted tree seedlings along a randomly located transect pattern in each plot. They returned in 2005, after nine of the plots had been logged, and counted the tree seedlings along the same transects. (Data from D.C. Donato et al., 2006. “Post-Wildfire Logging Hinders Regeneration and Increases Fire Risk,” Science, 311: 352.) The numbers of seedlings in the logged (L) and unlogged (U) plots are shown in Display 4.15.

 

 

 

 

 

Plot		Action		Seedlings2004		Seedlings2005		PercentLost
Plot 1		L		298		164		45.0
Plot 2		L		471		221		53.1
Plot 3		L		767		454		40.8
Plot 4		L		576		141		75.5
Plot 5		L		407		217		46.7

 

 

Analyze the data to see whether logging has any effect on the distribution of percentage of seedlings lost between 2004 and 2005, possibly using the following suggestions: (a) Use the rank- sum procedure to test for a difference between logged and unlogged plots. Also use the procedure to construct a 95% confidence interval on the difference in medians. (b) Use the t -tools to test for differences in mean percentages lost and to construct a 95% confidence interval. Compare the results with those in (a).