A reliable method of monitoring the numbers of wetland grazers is needed to assess whether the objective of increasing reedbuck, waterbuck, lechwe and puku populations is being achieved. Present air survey techniques are not suited to precise or accurate estimates of reedbuck, waterbuck, lechwe and puku populations. It might be thought that when lechwe are present in large numbers, air survey estimates would improve. However, because the animals are grouped in large herds, a new set of statistical problems arise as in the case of buffalo (Martin 2002b). As a technical problem, there are three key features that should guide the design of a monitoring system for the wetland grazers

1. hese four species should be treated as a unique group to be surveyed and not included in more general surveys which attempt to count all species;
2. Because the species are strictly confined to floodplain habitats (with the possible extension into adjacent riverine woodlands in the case of waterbuck), the survey can be restricted to a narrow belt along the main rivers;
3. Whatever survey technique is adopted, it needs to be repeatable over a number of years, not dependent on particular observers and not disproportionately expensive.
   

• Child and von Richter (1969) monitored waterbuck, lechwe and puku populations along the south bank of the Chobe river over a period of three years using simple routine counts from a vehicle travelling through the floodplain habitats roughly once every three days. The frequency of counts need not necessarily be as high as this but it is important that sufficient observations are collected to be statistically meaningful. This technique will not work in areas where there are no roads or where roads are impassable for much of the year.
• A line transect system in the floodplain habitats could have wide application in both conservancies and State protected areas. The transects need not necessarily be straight lines provided the route is repeatable. The transects should cut across the floodplain catena rather than follow the river in order to accommodate the habitat niches of the four species (Figure 3). A strong advantage of such a system is that it is not dependent on roads and can be implemented in the wet seasons.
• A low-level (100ft) air sample survey carried out annually (or more frequently) by a slow flying aircraft (e.g. Piper Supercub) along the floodplains is an attractive option. The simplest form of such a survey would be a single pass following the line of river but this would not necessarily produce the best estimates since animals far away from the river are likely to be missed. The floodplain habitats along the Kavango and the Kwando are up 10km wide in many places so that the next step up in sophistication would be to run transect lines perpendicular to the river. Finally, the standard of vegetation mapping in the floodplain habitats is so high (Mendelsohn and Roberts 1997) that it might even be possible to stratify the survey according to floodplain and riverine vegetation types - which would yield not only numbers but additional information on habitat occupancy.

Aerial surveys

Aerial survey data on reedbuck, waterbuck, lechwe and puku population sizes in Namibia are scant. On aerial surveys only a fraction of the true reedbuck population is likely to be seen because most of the animals are in cover and constitute the "invisible population". Because of their size, waterbuck are more likely to be seen than the other three species but when their numbers are low they tend to be missed by observers. When lechwe are abundant, the large herds are very visible from the air in open floodplain habitats. However, a different problem arises here - that of estimating numbers, which may be in the thousands, on a single pass of an aircraft flying at a speed of 100 knots 300 feet above the ground. When lechwe numbers are low, animals in small groups tend to be missed. Puku have not been recorded on any of the Namibian surveys in the Caprivi since their inception, although there are a number of sightings of them on the ground. This may be because of their low densities and because of their resemblance to other species such as impala, with which they frequently associate.

Law enforcement effort and illegal activity needs to be monitored to ensure that the objective for illegal hunting is being achieved. The present monitoring systems which are in place in the conservancies are adequate for this purpose.