a. UMEMENL os a —— 
THE USE OF LARGE SCALE AERIAL PHOTOGRAPHY TO DETERMINE 
WOOD BIOMASS IN THE ARID AND SEMI-ARID AREAS OF KENYA 
by 
H. Epp and D. Peden 
KENYA RANGELAND ECOLOGICAL MONITORING UNIT 
MINISTRY OF ENVIRONMENT AND NATURAL RESOURCES 
P.O. Box 47146 
NAIROBI, KENYA 
D. Herlocker 
UNESCO INTEGRATED PROJECT IN ARID LANDS 
P.O. Box 30592 
NAIROBI, KENYA 
ABSTRACT 
A multi-stage technique was developed for inventorying wood biomass in 
arid and semi-arid regions of Kenya. Previous studies using destructive 
sampling provided regressions relating tree wood biomass tc tree crown dia- 
meter. Ground-based measurements of tree crown diameter were related to 
crown diameter as measured on large scale, black and white aerial photographs. 
A statistical procedure was developed by which the two sets of regressions 
could be combined in order to predict the wood biomass and its standard error 
from photographic measurements of crown diameter. The method was applied to 
an example study area. The results indicate that the method provides a fast, 
cheap and adequately precise method of inventorying biomass over large areas. 
INTRODUCTION 
Up to 75% of all energy in Kenya is derived from wood fuel for which 
there is an increasing demand due to the increasing cost of imported petroleum 
products. At present, the high potential areas of Kenya, which cover only 18% 
of the country (Pratt & Gwynne, 1977) supply a large portion of the nations 
wood fuel. However, due to an increasing population, the demand for agricul- 
tural land can be expected to decrease the availability of high potential 
land for forest plantations. Kenya must therefore look to its arid and semi- 
arid areas, which cover 82% of the country, for the remainder of its fuel 
wood requirements. The low unit value of the woody material in this area 
requires inventory methods which are extensive, simple and fast. The very 
large area involved and its difficulty of access, especially during the rains, 
require an emphasis on aerial rather than ground based surveys. 
Extensive areas of forest and woodland can be quickly and cheaply 
surveyed from the air using large scale aerial photography and multi-stage 
sampling techniques (Spurr, 1948; Husch, 1964; Loetsch & Haller, 1964). 
Although generally less accurate than ground based surveys, they are far more 
efficient and cost effective (Aldrich, 1979). 
This paper describes a multi-stage survey technique using regression 
equations to predict tree biomass from crown diameter as measured on large 
scale photographs. These are developed through destructive harvesting of 
trees on a small number of plots and are applied through measurements of tree 
crown diameter on a much larger number of 70 mm and/or 35 mm aerial photo 
plots distributed throughout the sample area. 
415