Geographically, Donga lies within latitude 70'43''N of the equator and longitude 10E of the meridian with the land area of approximately 58 square kilometers. It is located in the in the southern Taraba, bounded on the Northeast by kurmi Local Government Area, on the Northwest by Bali Local Government, to the Southwest Wukari Local Government Area, while to the Southeast by Takum Local Government Area. Donga has ten (10) political wards namely: Fada, GyataAure, Gndindutse, Mararaba, Kadarko, Akate, Suntai, Asibity, Gayama and Nyita Ward. See Figure 1 and Figure 2 for location of the study area [10] .

The area has two marked seasons, the dry and the rainy season. The dry season starts from November to Mach

while the rainy season from April to October. The temperature is generally high with mean annual temperature at 32˚C recorded in April and low temperature can be as low as 10˚C experienced in December January, but the mean minimum annual temperature is 22˚C [10] .

Donga is underlain by sedimentary rock of post Cambrian land era, and mainly consists of gentle undulating land with plains. It has a river and the basin constitutes one of the fertile arable lands of the mid region of Nigeria. The hilly undulating land stands at an altitude of 120 - 130 meters above sea level [10] .

Donga Local Government falls within the woodland savannah zone a distinct type of vegetation, which is a transition between forest and savannah vegetation type. Thus there are both varieties of forest and savannah vegeta- tion in which natural plant cover consist of trees, shrub and grasses, some of them loose there leaves for only short period in the year. A forest variety such as oil palm is formed scattered all over the area [10] .

Donga Local Government Area has diverse ethnic groups which include Chamba, Ichen, Kpanzun, Hausa, Tiv and Fulani. According to 2006 Census Figure released by National Population Commission (NPC), Donga Local Government Area amounted to population of about 134,111 people [10] .

Considering the nature of the study, the size of the population as well the resources available for this research, the purposeful sampling method was used to divide the study area into identifiable units for the purpose of selecting samples. The local government has 10 wards and two wards were purposely sampled for this study. The reason was that the two council wards are more endemic in times of flooding, than any other council ward in the area. The two wards are: Gyata Aure and Fada wards.

A representation sample with known confidence and risk level was selected, based on the work of [11] . Based on this, 95% confidence level and an error limit of 10% were adopted.

The formula used by Yamane [11] goes thus:

where

n = required responses.

e² = error limit.

N = sample size.

We arrived at 100 questionnaires. To account for possible attrition, the number of subjects was increased to 130. The systematic random sampling was used in the administration of 130 questionnaires in the sampled wards. A total of 121 questionnaires were retrieved out of a total of 130 administered.

Percentage

where:

P = Percentage.

F = Frequency to be converted to percentage.

N = Numbers of frequencies.

Data obtained was analyzed via statistical (Spear Man rank correlation analysis) and descriptive statistics such as percentages and tables.

To examine public awareness and perception of flooding in Donga town, Donga local Government Area of Taraba State, the following hypotheses were postulated:

H0: there is no significant relationship between respondent’s educational status and their level of awareness of flooding in Donga town.

H0: There is no significant relationship between respondent’s perception of flooding and their level of education in Donga town.

This analysis was done with the aid of the Statistical Package for Social Sciences (SPSS).

The demographic data shows that 95% of the sampled respondents are male, while 5% are female as shown in Table 1. This disparity in gender proportion could be due to religious reasons as the study area is predominantly inhabited by Muslims. Under this religious understanding women are difficult to come by as Islam prohibits married women interacting freely with men.

Table 2 shows that 2.5% of the respondents are within the ages of 10 - 20 years, 28.1% were between 21 - 30 years, 31.4% between 31 - 40 years, 19% between 41 - 50 years and 19% 50 years and above. This corroborates the findings of [12] in Jalingo, which asserts that majority of Jalingo dwellers were of the age bracket 24 - 40 years.

Table 3 shows that 33% of the respondents attended only primary school, 29% attended secondary School, 21% attended tertiary institution. 2% had informal education and 4% vocational education. This confirms the findings of [13] which are supportive of the fact that a large proportion of Makurdi dwellers attended primary, secondary and tertiary institutions.

Table 4 shows the occupational status of the respondent. 26% of the respondents are engaged in fishing and farming, 18% are engaged in Trading/ Business, 16% are Artisans/craftsmen, and 13% are Civil servants. Based on field observation, most of the respondents are engaged in two (2) occupations. This confirms the finding of [12] which asserts that majority of the inhabitants of Jalingo metropolis are Farmers and Civil Servants.

Table 5 indicates that 11.6% are slightly aware of flooding, 88.4% are strongly aware of flooding in the area. This finding corroborates that of [12] which is supportive of the fact that majority of floodplain dwellers in Jalingo were aware of flooding before moving into the area, some were not.

The question sought to ascertain how long the respondents have lived in the study area for those who have lived long will appreciate the flood problem better. This is to ensure that the respondent have a reasonable idea of the extent of flooding. 19% have lived between 1 - 10 years, 22.3% 11 - 20 years, 22.3% 21 - 30 years, 19% 31 - 40 year, 8.3% 41 - 50 years and 9.1% 51 years and above. These findings are similar to that of [12] which is supportive of the fact that inhabitants of Jalingo floodplain have reasonable idea of the flooding problem in Jalingo metropolis because they have lived for some time in the floodplain (Table 6).

Table 7 shows that there are many causes of flooding in the study area. Rainfall accounts for 46.2% of the cause of flooding, proximity to the main river accounts for 45.5% and low land area accounts for 8.3% of the cause of flooding. This finding is similar to that of [14] , who conducted a flood risk assessment and management in Niger Delta Region of Nigeria. He noted that flood can be caused by heavy rainfall, highly accelerated snowmelt, severe winds over water, unusual high tide, tsunamis or failure of dams, levees, retention ponds or other structures that

Source: field survey 2015.

Source: field survey 2015.

Source: field survey 2015.

Source: field survey 2015.

Source: field survey 2015.

Source: field survey 2015.

Source: field survey 2015.

retained the water. Flooding can be exacerbated by increased amounts of impervious surface or by other natural hazards such as wildfires which reduce the supply of vegetation that can absorb rainfall. [15] cited in: [16] observes that rainfall intensity, duration and amount are generally believed to be the principal factors in most flood events in the tropics which are partly or wholly climatological in nature. Other factors only aided the incidence of flooding in urban areas of Nigeria. This has been confirmed by several studies cited in [16] - [19] .

H0: There is no significant relationship between respondent’s educational status and their level of awareness of flooding in Donga.

Statistical analyses using spear Man correlation coefficient analysis indicates that the calculated value (r calculated) is 0.051 and the critical value is 0.950 at 0.05 level of significance and 2 degree of freedom. Hence, we accept the null hypothesis which states that there is no significant relationship between respondent’s educational status and their level of awareness of flooding in Donga town. It means that educational attainment of respondents does not influence respondent’s awareness of flooding in the study area. The finding is similar to that of [20] .

H0: There is no significant relationship between respondent’s perception of flooding and respondents level of education in Donga town.

Statistical analyses using spear Man rank correlation coefficient analysis shows that the calculated value (r calculated) is 0.890 and the critical value is 0.950 at 0.05 level of significance and 2 degree of freedom. Hence, we accept the null hypothesis which states that there is no significant relationship between respondent’s perception of flooding and respondent’s educational status in Donga town. The finding is similar to that of [20] . This implies that in Donga town people’s perception of flooding is influenced by factors other than literacy level.