## 2.1 Description of Experimental Site

The experimental site (Fig. 1) for this study is located at the Teaching and Research farm (7.5o N, 4.5o E), Obafemi Awolowo University campus, Ile-Ife, Southwestern Nigeria. The measurement area is of dimension 18 m by 18 m and its surface was covered with blow-level grass which is cut periodically. Although, the dimension of the experimental site was not wide enough to allow for the required fetch necessary for effective flow measurement yet the location has the advantage of not being surrounded by trees and buildings thus obstructing surface flow and the surface roughness parameters greatly minimized. The field measurements took place for two months (June and July 2016). The climate of the study area is generally characterized by two seasons: namely the wet and dry seasons with the wet season spanning between March to October and the dry season lasting from November to February. According to (Jegede et al. 2004), the variation of this season is a function

**<Insert** Fig. 1 **>**

of the meridional movement of the Inter-Tropical Discontinuity (ITD) which demarcates the warm and moist South-Westerly trade wind at the surface from the hot and dry North-Easterly trade winds (Jegede et.al., 2006). During May/June, which is the onset of the wet season at Ile-Ife, it is within weather zone B (which extends 200–400 km south of the surface position of the ITD). The zone is characterized by suppressed convection resulting in cumulus clouds and precipitation is limited to light showers, whereas in August/September, which is at the peak of the wet season in the area, Ile-Ife falls within the weather zone D characterized by stratus cloud and is accompanied by light rains and drizzles with occasional moderate thunderstorm activities (Ayoola et.al., 2014).

## 2.2 Instrumentation and Data Processing

Multilevel measurements of wind speed, air temperature, and moisture were conducted by installing meteorological sensors on a 15 m mast at the experimental site. The cup anemometers (model A100L2) and air temperature/relative humidity sensors (HMP60) were arranged in a log-linear position at five (5) heights of 1.14 m, 1.88 m, 3.30 m, 6.30 m, and 12.10 m on the same mast (Fig. 2). Simultaneously with the profile measurements, an eddy covariance (EC) system comprising of a 3D ultrasonic anemometer (CSAT3) and an open path infrared gas analyzer (LI-7500) placed on a mast of the height of 1.81 m was co-located to measure the turbulent fluxes of momentum, sensible and latent heat, and stability parameters at the surface. Data acquisition and storage were controlled by the use of two programmable CR1000 data loggers. The profile measurements were averaged and stored at One (1) minute values while the turbulent fluxes were recorded at 10 Hz.

**<Insert** Fig. 2 **>**

The raw data were reduced to Thirty (30) minutes averages of mean wind speed, air temperature, and relative humidity for each day. The log-linear curve in equations 1, and 2 were fitted to the estimated profile of air potential temperature, wind speed, and air-specific humidity data using MATLAB.

$$\stackrel{-}{\theta }\left(z\right)=az+blnz+c$$

1

$$\stackrel{-}{q}\left(z\right)={a}^{{\prime }{\prime }}z+{b}^{{\prime }{\prime }}lnz+{c}^{{\prime }{\prime }}$$

2

From these fits, a gradient of the resulting profile was obtained for potential temperature and specific humidity given by equations 3 and 4 respectively:

\(\frac{\partial \stackrel{-}{\theta }}{\partial z}=a+\frac{b}{z}\) (3) \(\) \(\frac{\partial \stackrel{-}{q}}{\partial z}={a}^{{\prime }}+\frac{{b}^{{\prime }}}{z}\) (4)

In this study, the widely applied expression of Businger *et al.* (1971) has been used to obtain\({ {\phi }}_{h}\left(\xi \right)\), Eq. (5) which is essentially equivalent to\({ {\phi }}_{q}\left(\xi \right)\).

The expression is given as:

$${ \phi }_{h}=\begin{array}{c}0.74{\left(1-9\xi \right)}^{-\frac{1}{2}} \left(unstable condition\right)\\ 0.74+4.7\xi \left(stable condition\right) \end{array}$$

5

Based on the Monin-Obukhov similarity theory for the horizontally homogenous and stationary surface layer, the non-dimensional profile functions of air potential temperature and air-specific humidity are expressed as equations 6 and 7:

$${ {\phi }}_{h}\left(\xi \right)=\frac{kz}{{\theta }_{*}}\frac{\partial \stackrel{-}{\theta }}{\partial z}$$

6

$${ {\phi }}_{q}\left(\xi \right)=\frac{kz}{{q}_{*}}\frac{\partial \stackrel{-}{q}}{\partial z}$$

7

where\(\stackrel{-}{\theta }\)and\(\stackrel{-}{q }\)are mean wind speed, potential temperature, and specific humidity respectively, \(k\) is the von Karman constant which is a universal constant independent of flow or surface characteristics and has been determined by many researchers throughout the last decades (see Högström, 1988). Today, \(0.40\) is the accepted value for \(k,\) \(z\) is the measurement height, \({u}_{*}\) friction velocity (related to turbulent momentum flux), \({\theta }_{*}\) the scale temperature (related to turbulent heat flux), \({q}_{*}\) the scale humidity (related to the latent heat flux), and\(\xi\)=z/L is a stability parameter defined as the ratio of height, \(z\), to the Obukhov length, \(L\), (characteristic height in the surface layer).

With the gradients and the non-dimensional similarity functions of heat and momentum\({ {\phi }}_{h}\left(\xi \right)\), \({{\phi }}_{q}\left(\xi \right)\)obtained;\({ \theta }_{*}, { and q}_{*}\) were then determined from equations 6 and 7 above.

From the estimated \({\theta }_{*}\)and\({ q}_{*}\)following the aforementioned steps, the turbulent fluxes of sensible, \({H}_{s},\) and latent heat, \({H}_{L},\) were obtained from expressions given in equations 8 and 9:

$${ H}_{s}={-\rho C}_{p}{u}_{*}{\theta }_{*}$$

8

$${ H}_{L}=\rho \lambda {q}_{*}{u}_{*}$$

9

where \(\rho\)is the air density, \({C}_{p}\) is the specific heat capacity of heat at constant pressure, \(\lambda\)is the latent heat of vaporization, \({u}_{* }\)is the friction velocity, \({\theta }_{*}\)is the scale temperature and\({ q}_{*}\)is the scale humidity as earlier discussed. The computed fluxes were then compared with the direct measurement obtained from the EC system.