Epsilon Archive for Student Projects

Abstract

This report is an assessment of the feasibility of using aerial photographs to estimate number of stems per hectare in young forest stands, used together with upper height to determine the need for pre-commercial thinning or ordinary thinning cutting. Aerial photo-interpretation was conducted in 2007 on assignment of Stora Enso Skog AB. A mean height function based on stand age produced with data from Bergvik Skog AB’s 2006 forest inventory assessment was used as a proxy for upper height and used together with photo-interpreted number of stems per hectare to determine the need for thinning cutting according to Bergvik Skog AB’s definitions. The test area is located north of Falun on Bergvik Skog AB’s land.

Flight altitude was approximately 900 m above ground and the aerial photos were taken by L&L Flygbildteknik AB. The aerial photos were taken with a Hasselblad 555eld camera using a digital shutter PhaseOne P45+ and a 40 mm lens from a Cessna 172. The selection of flight altitude was based on experiences from a pilot study made in 2007.

A field inventory was conducted during the spring and summer of 2008 in 29 stands with a total area of 417 ha productive forest. Stand variables measured were: stems per hectare, site index and upper height. Each variable was first measured on 7 – 16 circular plots per stand depending on the stand size. The plot radius was set to 5 or 7 m depending on the number of stems per hectare in the stand. Mean values were then calculated for each stand.

Results show that using aerial photographs to estimate stems per hectare lead to systematic underestimation of the true stem numbers. For stems ≥ 1.3 m in height the average difference between photo-interpreted and field measured values was about 1600 stems per hectare. For stems ≥ 6 cm in diameter at 1.3 m height, the difference between photo-interpreted and field measured stem numbers was about 450 stems per hectare.

Comparisons between the mean height from Bergvik Skog AB’s mean height function and upper height from the field inventory show that the height function underestimates the upper height in pine stands by 2.7 m and spruce stands by 1.1 m.

The fact that the mean height underestimates upper height and that stems per hectare were underestimated using photo-interpretation is also reflected in the stand treatments based on these two variables, as compared to treatments based on field data.

A regression function was made to estimate stems per hectare ≥ 6 cm in diameter (at 1.3 m height) at stand level using stems per hectare counted in the aerial photography and mean height as independent variables. Two regression functions were also created to estimate upper height using mean height from Bergvik Skog AB’s height function as independent variable, one for pine and one for spruce stands.

A new treatment classification was made for stands with thinning cutting potential using upper height and stems per hectare > 6 cm in diameter estimated with the above described regression functions. The result shows a classification accuracy of 74% compared to 63% obtained using photo-interpreted number of stems per hectare and mean height according to Bergvik Skog AB’s height function.