The seasonal pattern of rainfall in the Top End means that most of the streams do not flow all year round. Instead, these "seasonal streams" flow for just a few months during the wet season and remain dry for much of the year. During this dry period, fires frequently burn large parts of the stream catchment, but no one knows what effect this has on the streams themselves. As part of the Kapalga experiment, we studied the effect of catchment burning on the water quality and on the aquatic plants and animals in seasonal streams. Initially we focused on the Unburnt and Late treatments.

The seasonal pattern of rainfall in the Top End means that most of the streams do not flow all year round. Instead, these "seasonal streams" flow for just a few months during the wet season and remain dry for much of the year. During this dry period, fires frequently burn large parts of the stream catchment, but no one knows what effect this has on the streams themselves. As part of the Kapalga experiment, we studied the effect of catchment burning on the water quality and on the aquatic plants and animals in seasonal streams. Initially we focused on the Unburnt and Late treatments.

Riparian vegetation was very sensitive to Late fires. Compared with the Late catchments, Unburnt riparian zones had twice as many species and about three times the density of woody plants. Vines and climbers were also much more abundant and diverse in the Unburnt riparian zones. The most common eucalypt along these streams, Eucalyptus alba, set seed in the Unburnt catchments but didn't even flower once they had been burnt Late. In short, Late burning seemed to either kill the woody plants and vines or reduced reproduction; clearly a detrimental effect. Within the stream itself, however, it was a very different story.

There were very few differences in any of the water quality features that were measured. Twice as much sediment was exported from the Late catchment than the Unburnt catchment. Unlike the Unburnt streams, Late streams had several episodic runoff events, of "poor" water quality, before continuous wet season flow.

Probably the most obvious life in these streams was the lilies, wild rice and other aquatic plants that were common in pools at the end of the wet season. There was a dramatic difference between aquatic plants in Late and Unburnt catchments. Streams in Late catchments contained six times as many different species and over ten times the biomass of aquatic plants. In fact, it was difficult to find any aquatic plants in most of the Unburnt catchments. Less obvious than the water plants, are the multitudes of aquatic invertebrates that inhabit these streams. Like the aquatic plants, these tiny animals also seemed to benefit from catchment burning. Throughout the wet season, streams in burnt catchments had 50-100% more species of aquatic invertebrates than streams from Unburnt catchments. At some times of year there was also a greater overall abundance of aquatic invertebrates.

Research at Kapalga over the last three years indicates that compared with Late and Unburnt regimes, burning early in the dry season results in intermediate riparian tree density, riparian canopy cover and the biomass of aquatic plants. Catchments burnt early in the dry season had similar water quality, similarly high richness of riparian vegetation, and similarly low richness of aquatic plants to that found in Unburnt catchments. Burning early in the dry season therefore seems to be a reasonable trade-off between maximising the benefits for riparian vegetation and aquatic biota. Regardless of the fire management regime used, our research has shown that there is a strong link between tropical savannas streams and their catchments. Consequently, fire management, which has been traditionally considered a "land" management issue, has clear implications for the management of riparian and aquatic resources.