Experimental investigation of turbulent bursts in the atmospheric surface layer over the tropics

Abstract

Turbulence, an important characteristic of the atmospheric boundary layer (ABL), is mainly responsible for the transport of heat, moisture, momentum, pollutants, etc., in the atmosphere. As with any turbulent flow, our understanding of the ABL has to gain in large part from observational studies.

So far, several observational experiments have been carried out in the middle latitudes, but similar experiments in the tropical regions have been rare. This thesis describes an observational study of the turbulent flow in the atmospheric surface layer of the Indian tropical region, and includes accounts of various stages in conducting the field experiments, the detailed quality checks made to validate the acquired data, and finally, a description of the turbulent flow in terms of “events,” which has been considered in the recent past as possibly a more natural way of description.

As a part of the national experiment called MONTBLEX-90 (standing for Monsoon Trough Boundary Layer Experiment 1990), surface layer observations have been carried out during the monsoon period of the year 1990 using four 30m instrumented towers, one each at Kharagpur, Varanasi, Jodhpur, and Delhi, representing regions of deep moist convection, shallow dry convection, and the transition between the two. Sensors for measuring both mean and fluctuating quantities of wind speed, temperature, and humidity were used, with the majority of them indigenously built. A PC-based data acquisition system built by us at the Centre for Atmospheric Sciences, IISc, was used to acquire the fast-response data, and an imported data logger was used while the fluctuating quantities were acquired for 10 to 15 minutes every hour or once in three hours, depending on whether there were any synoptic-scale disturbances.

It was a great experience when we were out in the field trying to record nature’s play through sensors and other electronic instrumentation, and a glimpse of the observational phase is also provided in the thesis. A large set of data was acquired for a duration of about three months, covering different stability conditions and different phases of the monsoon. Several quality checks were made on randomly selected data to assess the quality of the data. On the whole, it turned out that observations at Jodhpur were most successful in terms of the number of good sets of continuous data, so we begin our main investigation of events (or bursts) in the flow using data from that site.

Employing the variable interval time averaging (VITA) technique on the fluctuating components of horizontal streamwise velocity (u), vertical velocity (w), and the temperature (T) derived from sonic anemometer data, we first compare the burst characteristics in Jodhpur with earlier work on mid-latitude data (Boulder, USA), both taken when the stability was near-neutral. Many characteristics of the events were found to be similar between the two sites, and we confirmed the occurrence of the most intense events in u at longer averaging times and those in w at shorter averaging times.

The presence of both accelerating and decelerating types of events was confirmed, and the need to treat them separately has been highlighted. In an attempt to identify the events with objectively determined intensity and duration parameters close to their natural values in the flow, we have improved the event detection algorithm taking advantage of certain basic properties of the VITA technique. Using the modified algorithm, we are able to draw a chronicle of events, each with appropriate intensity and duration.

Based on this chronicle, we have seen the characteristic signature of both accelerating and decelerating events, which are found to be very similar to those found in earlier work on both laboratory and atmospheric flows. With the large set of data files available, various event characteristics have been derived for the first time under different stability conditions, such as neutral, stable, and unstable. The conditional average of u and T signals during eventful times of the u signal describes closely the prevailing stability of the atmospheric surface layer. The contribution from events to the energy (u², w², T²) under different stability conditions is also discussed.

The drawbacks of the VITA technique for flux data have been discussed in some detail, providing confirmation of the results of some earlier studies. From quadrant plots of momentum and heat flux data, the duration of occupation of each quadrant, their contributions to total flux, and their variation with stability are seen to provide interesting information and patterns. Also, the productive, counter-productive, and idle periods are identified and later described in terms of events. A simple threshold technique is then described for the detection of events in the flux data. Momentum and heat flux data are analyzed based on this technique, again for different stability conditions, and a chronicle of positive and negative events is determined.

Based on such chronicles, an ensemble average of events is computed. Interestingly, both positive and negative events exhibit a coherent structure. The burstiness parameter, introduced by Basu, Haque, and Kailas (1990), when computed with the events detected from the present technique, is found to have much higher values (greater than 0.70 for most of the data considered). Momentum flux events are seen to have higher burstiness when compared to heat flux events.

The coincidence coefficient, defined for momentum and heat flux events, is seen to be very high, in the range 70–90%. The diurnal variation of various characteristics for momentum and heat flux events is sketched using hourly data sets for one full day. The major conclusion from this study is that the dominance of positive (or negative) events changes clearly with the time of day (or stability) in the case of heat flux events, while in the case of momentum flux, positive events dominate during most of the day, except during periods of strong instability (namely in the afternoons), when negative events come close to positive events in all respects, occasionally even dominating them.