Empirical hydrology data (soil moisture, surface runoff, stream depth and groundwater) is obtained from an experiment site in the Tarland sub-catchment of the River Dee in Aberdeenshire, Scotland. This data is used to categorise the riparian buffer behavioural response to storm events, the surrounding landscape and its environmental characteristics. These categories are statistically analysed to clarify the field scale ability of riparian buffers to attenuate runoff. A threshold of effectiveness is assessed by comparing the characteristics and behaviour of: rainfall events, environmental conditions and riparian runoff, as well as an examination of interactions between the hydrological variables. Subsequently, the Soil and Water Assessment Tool (SWAT) hydrological model is utilised to assess the catchment-wide implementation of riparian buffers, overlaying the hydrological properties of the buffers in the catchment and using the field data as calibration.
RESULTS & DISCUSSION
Figure 1. For a rainfall event in February 2014, the top graph shows rainfall and stream depth and the lower graph shows soil moisture and surface runoff. The variables are all displayed for the period: 01:00am (5th Feb) to 05:30am (6th Feb), a total of 28.5 hours.
Figure 1 demonstrates how the riparian buffer behaves during a rainfall event where the key characteristics are: wet antecedent conditions, bare soil in the adjacent field, limited vegetation growth and a prolonged heavy rainfall event. The behaviour of the riparian buffer is evident from figure 1 whereby there is:
* an insignificant fluctuation in stream depth and soil moisture inside the riparian buffer (INRIP);
* no interaction between rainfall and INRIP surface runoff;
* a slight response of the soil moisture outside the buffer strip (OUTRIP), which is often found to reach a 'plateau' state before declining;
* and finally, the OUTRIP surface runoff shows that it engages with the rainfall event and has higher surface runoff than INRIP.
These behaviours suggest the riparian buffer is attenuating runoff and therefore reducing flood risk locally. Under similar conditions in January 2014, the same response (as seen in figure 1) is evident. However, further data analysis of subsequent months is necessary because:
* There is some evidence that the riparian buffer shows a very different behaviour in response to varying environmental conditions.
* INRIP soil is consistently wetter than the OUTRIP soil, a behaviour that is likely to be due to hyporheic exchange.
* The soil moisture 'plateau' requires a determination of the field capacity of the soil and comparing the plateaus over time to distinguish behaviours and clarify the cause of this.
* The stream flow values require further investigation as fluctuations of water level are minimal.
With further data collection in the coming months, statistical analysis will be carried out and data will be suitable for model calibration. Further modelling work will assess the catchment scale impact.
Initial results of the riparian buffer strip hydrology data suggest that in some instances the riparian buffer performs as expected however, there are some data demonstrating a contrasting behaviour whereby surface runoff is higher inside the riparian buffer. Further data collection in conjunction with continued classification of riparian buffer behaviour responses and landscape/ environmental characteristics is necessary to elucidate the ability of riparian buffer strips to reduce flood risk. Moreover, with continued data collection, hydrological modelling can simultaneously address the challenges of riparian buffer's optimal width, thresholds of effectiveness and scale of implementation. Both the field data and model data will contribute to a scarce evidence base on riparian buffer NFM effectiveness.