Abstract:
: This study sought to establish an efficient inventory protocol to estimate regeneration
stock and dynamics in natural forests. Field and computer outputs were integrated to develop
complete inventory protocol for selected natural forest types in Kenya. Inventory cost and
precision for four plot sizes (5 m × 5 m, 10 m × 5 m, 10 m × 10 m and 20 m × 20 m) were
determined and compared. Specific objectives were to determine (i) precision levels of estimating
tree seedling density using different plot sizes across forest types; and (ii) optimum plot size which
minimise both sampling error and inventory effort for use in each forest type. Seedling counts and
time taken per plot were recorded systematically over 400, 200, 100 and 25 plots ha-1
for the
respective plot sizes. Larger plots and their data were created by merging smaller ones through
programming with R Software. Smallest population mean-variance,
was obtained using
data from 5 m x 5 m plots. Both precision and inventory effort varied with plot size used, but in
reverse directions. Seedlings population mean errors were 15.4% of the mean for rain forest,
14.7% for moist montane forest and 9.9 % for dry forest type. Inventory cost decreased with
increasing data compilation unit size, e.g., 50.42 hrs ha-1
for 25 m2 unit to 3.21 hrs ha-1
for 400 m2
unit in rainforest. Similar trend was observed in other forest types. Recommended plot sizes for
tree seedlings are 75 m2
; 62.4 m2 and 88.4 m2 for Kakamega rain forest, Mt Elgon montane forest
and Loruk dry woodland forest, respectively. These plot sizes gave acceptable uncertainty levels
between ±11% and ±17% of mean estimate ha-1
. Tree diameter distributions from 5 m x 5 m plots
revealed that tree component recruitment was irregular over time across forest ty
