This survey was based on a survey effort 924 trap-nights. They split up unequally over the study regions: in Kasanka NP 748 trap-nights were performed. A comprehensive survey requires a survey effort of more than 1,000 trap-nights (GIMAN et al. 2007) not reached by this study. Furthermore, the camera trap stations did not cover the entire study regions but rather specific spots. The choice of the spots in Kasanka NP was a result of the reflection about where captures of predators might be more possible or more probable than in other areas and was based among others on information of FRANK WILLEMS. Hence, a considerable amount of camera traps was mounted at Fibwe. Placed further as a complement to the line transects in Kasanka NP, the information gathered by the camera traps were comparable to the results of line transect sampling. In the study regions, results have to be considered as trends or additional information, not as a stand-alone survey. However, photos collected by less statistically sampling designs may still provide useful data (JENKS et al. 2011).
Especially for the abundance of non-individually identifiable species, indices can be problematic because a distinction between cases of single individuals with numerous records and cases of numerous individuals with single records is difficult (TROLLE et al. 2008).
Therefore, a determination of independent captures as done by O’BRIEN et al. (2003) and as performed for this study is very important, also in order to reduce the risk of overestimation
mentioned by SILVEIRA et al. (2003). The RAI does not give actual numbers or densities in the study areas, but relative abundances; however, it facilitates comparison (JENKS et al. 2011).
Further, this relative abundance can be compared amongst animals in the study regions.
According to the RAI, the puku was the most abundant antelope in Kasanka NP. This agreed with the observations of the line transect survey, the resulting observations and estimated densities. Contrastingly for bushbuck: this antelope was ranked second and half as abundant as puku according to the RAI, but reached only a fraction of the puku’s densities according to the data from the line transect survey. The RAI underlines that bushbuck were common in Kasanka NP. This might be similar but less pronounced for bush duiker. During line transect sampling, reedbuck and hartebeest were detected, but their abundance was calculated low. Although observed during line transect sampling, sable and orebi were not captured by the camera-traps. None of the antelope species was observed by camera-traps only.
In PPKR, as a supplement to the observations during line transect sampling, the camera-traps captured the sable antelope. Hartebeest, reedbuck and grysbok were not photographed, although recorded on transects. In order to capture hartebeest and reedbuck, the camera-traps should probably have covered rather grassland at a certain distance to Kafue River than the area close to its boarders, and it might be that this is the reason why none of these animals were photographed. However, the camera trap at the waterhole on the transect line Plains, where hartebeest and reedbuck were seen during line transect sampling, did not capture these antelopes either. In PPKR, according to the RAI, the puku was the second most common antelope. Impala were most abundant in this region.
This was also the result of the transect line sampling. Concerning bushbuck, as in Kasanka NP, the results of the RAI, in contrast to the results of the line transects, might fit better to its actual abundance. Similarly, relative abundance between puku, waterbuck and greater kudu was less biased towards the puku as in the DISTANCE estimations. Results of the RAI might fit better to the actual situation; they suggest a higher relative abundance of these antelopes than the results from the line transects in the study region.
In general, detectability of species is higher for habitat generalist, terrestrial and large- bodied species (TREVES et al. 2010). This explains low RAI values of the small carnivores, but also of sitatunga. Further, this statement might explain the difference in RAI values of bushbuck and bush duiker, although bush duiker might not have been less abundant than bushbuck.
The RAI values of puku in different areas underlined the findings of the line transect surveys. Concerning RAI and estimated population densities, puku were very abundant at Puku-Loop. The area at Kabwe, partly covering the northern end of Puku-Loop transect line, resulted in a high abundance of puku as well. The area at Fibwe and Katwa provided medium densities of puku; they were similar concerning RAI, but based on the line transects, Fibwe had a higher density of puku. This might be caused by the position of camera traps and transect line: while the transect lines led through Fibwe plain, most of the camera trap stations were in or at the border of miombo woodlands. Puku did occur here, but at lower abundance than on grassland. Camera traps revealed a low occurrence of puku at Chikufwe as did the observations of the line transect. However, they both might indicate a rather vagrant status of puku on that site. The rather dense evergreen forest at Fibwe was not covered by the transect lines. Although occurring in miombo woodlands, puku were not detected in this dense evergreen forest.
Bushbuck were found at high abundance in the evergreen forest at Fibwe which underlines their general habitat preferences (LESLIE JR 2011b). Sitatunga were detected at the two focal areas at Fibwe. This area around Fibwe, characterised by a high amount of swamps, wetland areas and thickets matches with the habitat choice of this antelope (MAY &
LINDHOLM 2013). The western part of the park around Chikufwe provides less waterbodies.
Reedbuck and hartebeest were detected only at Chikufwe, similar to the observations on the line transects.
Although occurring at high abundance at the artificial waterhole near Kaingu Lodge, puku were never observed at the saltlick at about 50 m from the waterhole. Beside the observation during this study, puku were generally not seen at the salt (TOM HEINECKEN,pers.
comm.), but at the waterhole only. like bushbuck, while impala and kudu used both sites.
Impala, especially in the dry season, browse to a considerable amount, comparable to kudu (GAGNON & CHEW 2000), whereas the puku’s food is composed almost entirely of grasses (Chapter 3). Analyses of mineral content of the actual food plants might reveal more information. Nevertheless, puku’s food might be richer in minerals and sufficient for the puku’s requirements. However, these observations need to be intensified.
Generally, more cameras, installed for longer periods, placed at more stations and perhaps following a more standardised study design could provide more detailed information about the species and their abundances within the study areas. By this way, the occurrence of rare species might be proven and the relative abundance of all (larger mammal) species in the study regions could be assessed.