Atmospheric convection is sensitive to the nature of the surface and its temperature. Both dry (without cloud) and moist (with cloud) convections depend on the surface temperature. Surface temperature is of critical importance in several practical applications like human comfort and crop cultivation. In the climate change scenario too, variations in the surface temperature take the center stage. Therefore, prediction of surface temperature is important. The evolution of the temperature is governed by the energy equation and the surface temperature by the surface energy balance. Important components of the surface energy balance are radiation (incoming solar radiation, reflected solar radiation, incoming and outgoing longwave radiation), sensible and latent heat fluxes and heat flux into the ground (called ground heat flux). A large number of individual and collective observations have been carried out in the past to understand the atmospheric boundary layer and the surface energy budgets. However a major share of the observations is from mid-latitudes. There have been few experiments carried out in India, for example, MONTBLEX, LASPEX, etc. One common drawback among these experiments is that the data time series is discontinuous and continuous measurements covering an entire season are lacking. Moreover these measurements were not comprehensive and hence did not allowed to calculate complete surface energy balance – in some cases radiation data is not available while in some humidity data. Therefore, continuous time series of sufficient duration and covering all variables needed to look at the seasonal energy balance based on measurements alone is missing in the Indian context. New programmes with the main objective of predicting convection are being planned in India. For example, PROWNAM (Prediction of Regional Weather with Observational Meso-Network and Atmospheric Modeling) is aimed at predicting the short term weather at SHAR and STORM (Severe Thunderstorms – Observations and Regional Modeling) aims to predict the occurrence of severe thunderstorms in the northeastern India. In both these programmes, measurement of all components of surface energy balance is one of the main objectives. However, the minimum configuration and data accuracy requirements for the flux towers, sensitivity of computed fluxes on data accuracy have not been carefully evaluated. This thesis is aimed at filling this gap. As a part of my work, a 10 m high micrometeorological tower was installed in an open area within the Indian Institute of Science (IISc) Air Field. Temperature, relative humidity and wind speed and direction instruments were mounted at two levels, 2 m and 8 m. All components of radiation were measured. Data, sampled every 5 s and averaged for 2 minutes were continuously stored, starting May 2006 onwards. Soil temperature was measured at 4 depths, 5 cm, 10 cm, 15 cm and 20 cm. In addition, a sonic anemometer capable of measuring 3 components of velocity and air temperature was installed at 2 m height, and data was collected for more than a month to enable the calculation of momentum and buoyancy fluxes using the Eddy correlation method (ECM). The present work evaluated the sensitivity of the fluxes for small calibration errors and quantified the minimum data accuracies and configuration needed for flux measurement with the Profile method (PM). After applying corrections, the comparison of fluxes from PM and ECM are in good agreement. The complete long-term surface energy balances is calculated in terms of source and sink. One aspect that emerges from the observation is that the seasonal variation in the sink term is relatively small (150-170 Wm-2) whereas the source term shows much larger variation from 180-250 Wm-2. A method has been implemented by which the ground surface temperature can be estimated by using the subsurface temperature timeseries by the method of Fourier decomposition and using the Fourier heat conduction equation. In addition we can compute the thermal diffusivity of the soil by using the amplitude and phase information of the sub-surface soil time series. The estimated temperatures from this method and one that estimated from radiation method are in good agreement with the maximum difference being less than 0º C.