Spatio-Temporal Assessment of Vegetation Resource Dynamics in Nigeria from SPOT Satellite Imageries

Vegetation resources in Nigeria are of vital importance for the sustainable development of the country. However, this essential resource is in danger due to the effect of anthropogenic and climate induced impacts. Currently desert encroachment which cuts across the Sahel is affecting most of the states in the northern part of the country particularly the eleven states considered by the Federal Ministry of Environment in Nigeria as the frontline states. Several studies on the Nigerian environment have shown that there are serious threats to the general environment particularly vegetation. Due to population growth and the need for housing as well as the expansion of the over-utilised farmlands across these states, places considered as reserved areas across the country are being exploited to the detriment of the vegetal resources particularly the forest and rangeland areas. This study utilized Idrisi TerrSet (version 18) raster-based remote sensing and GIS software to analyse seventy two (72) dekadal Normalised Vegetation Index (NDVI) imageries from SPOT satellite covering Nigeria in order to assess the anthropogenic and likely climatic impacts on the vegetal resources using the forward t-mode Principal Component Analysis (PCA) with standardised principal components. Results indicated that Component 1 which explains about 69% of the 72 time-series NDVI imageries shows typical vegetation cover over the study area within the time period under study. While component two indicated a cyclic trend dif-ferentiating gration of socio-economic and high spatial resolution data into an assessment of this kind in future studies is encouraged.


Introduction
One of the vital resources across Nigeria being threatened by a combination of natural and man-made impacts is vegetation. Accordingly, the issue of desert encroachment particularly across the northern parts of the country referred to as the frontline states namely: Adamawa, Bauchi, Borno, Gombe, Jigawa, Kano, Katsina, Kebbi, Sokoto, Yobe and Zamfara states is a very serious issue of concern ( Figure 1(b)). Desertification in Nigeria is caused by land degradation in the northern states where poor land management and environmental pressure is on the increase [1] [2]. In the southern part of the country this is caused by mismanagement of the environment due to activities related to exploration of petroleum resources. All these affect the vegetation status across the country.
Vegetation influences the energy balance at the earth's surface and in the atmospheric boundary layer, often mitigating extremes of local climate [3]. The vegetation cover of Nigeria reflects past and present climatic variations and its current status particularly forest areas, had an annual deforestation rate of 5% per annum of its closed forests [4]. This can be considered as one of the highest in the world. Hence, rapid population growth and demands for economic development on a relatively natural, and in some areas, undisturbed vegetation which is not properly managed can also lead to permanent conversion of most vegetated areas to other forms of land use such as agriculture and housing in Nigeria. Studies dealing with vegetation dynamics can be undertaken using different remote sensing and GIS techniques. For example, [5] utilised remotely sensed data for assessing vegetation resources of Nigeria. Other studies which utilised NDVI from satellite data [6]- [11] have shown that NDVI data can provide an effective measure of photosynthetically active biomass. Other environmental studies relating to vegetation dynamics covering whole or part of Nigeria using satellite data include [12]- [18]. On the other hand, vegetation provides habitat to wildlife and ecosystem services such as food and fuel, timber, cash crops, pulp, fruits, robes, clothes and many game reserves [19]. Furthermore, the use of Principal Component Analysis (PCA) employing the standardised principal components (SPC) in environmental studies including vegetations resource dynamics using remotely sensed data analysed in a geographical information systems (GIS) environment indicated clearly that the major element of variability is that which occurs spatially [20] [21] [22]. However, the earlier studies on Nigeria did not utilise PCA with standardised principal components (SPC), hence the choice of this methodology using SPC. Principal Component Analysis using SPC allows any set of original satellite imageries or data to be transformed to a set of new images (referred to as components). Such component images in most cases contain all of the information in the original images entered into such time-series and, are uncorrelated with one another [23] [24]. The interpretation of the extracted component images thus, relies on a combination of spatial and temporal analysis. According to the United Nations [25], about 5 million hectares of degraded land across these states have been restored based on the review and progress report by Nigeria on the implementation of the Agenda 21 of the United Nations [26]. The purpose of this broad vegetation dynamic assessment therefore, is to verify this assertion and to present a cost effective approach where seventy two (72) remotely sense data derived from SPOT satellite in the form of dekadal Normalised Difference Vegetation Index (NDVI) ( Table 1) were analyzed within a GIS environment in order to highlight areas or serious concern with regards to changes in vegetation across Nigeria. In particular, more emphasis will be on the front-line states during the El-Nino and La-Nina (ENSO) periods of 1999 and 2009 respectively. When these areas of serious concern are identified either due to the impact of the ENSO or other anthropogenic factors, they can be targeted with systematic monitoring using a very high spatial resolution satellite imageries for sustainable vegetation resources management across the country.

Location and Geographical Setting
The study area covers mainly Nigeria but includes southern part of Niger Re- The climatic condition of the northern part of Nigeria exhibits only two distinct seasons, namely, short wet and a prolonged dry season. The seasonal pattern of climatic conditions over Nigeria gives rise to four seasons in the south and two in the north. This is the result of annual total rainfall occurrence and distribution, which is more predominant in the south than in the north. The mean annual rainfall along the coast in the south-east is about 4000 mm while it is slightly above 500 mm around Kano in the north, particularly during the El-Nińo phase of 2009 ( Figure 2(b)). However, the monthly rainfall during the La-Nińa phase of 1999 was above 600mm in one of these front-line states of Kebbi in Yelwa (Figure 2(a)). According to [27] however, the annual rainfall in the northern part of Nigeria study area (southern part of the Niger Republic and the extreme northern part of Nigeria) falling in the savanna vegetation zone of  The rainy season also varies, as it ranges from eight to ten months in the southern forest zones (towards the southern part of the study area) to less than four  [29]. The relationships between soil types and vegetation as well as between soil and the local topography can also be noticed for example, in areas of smooth relief. For example, in some parts of the north-west of the country around Kebbi state or in some parts of the north-east (around Maiduguri in Borno State) there are areas of smooth relief compared to areas around Adamawa state where the elevation is high, and the upper slopes mostly contain inactive soil, rather than the clayish type that can be seen in the immediate surroundings [30]. Furthermore, the soils on the lower slopes are formed by the washing down of material and are likely to be more stony and sandy, and hence, the vegetation around these 11 frontline states ( Figure 1(b)) is very scantly and susceptible to erosion.

Results and Discussion
This broad analysis of vegetation dynamics with PCA utilized the forward t-mode process where each image band was analyzed as a (temporal) variable which resulted into new set of principal component images that are uncorrelated with each other and explain progressively less of the variance found in the original set of NDVI imageries. A table of the component loadings and eigenvectors were also derived as part of the output. For t-mode forward PCA, using standardized variables utilized here, was meant for data compression or noise removal which essentially gives equal weight to all the bands and, it is often the case that the first four components typically explain close to or slightly 90 percent of the total variance in the original data set. Those components explaining less than a certain percent of the variance can be dropped and the inverse procedure used to reconstruct a new data set. Although results for this assessment yielded 72 components and their corresponding loading scores, the discussion is limited to only four component images and their corresponding loading scores as they represent 86% of the total variances from the original NDVI images utilized in the analysis, thereby indicating changes of varying degrees across Nigeria.

Component 1 Image and Its Loading Scores
Component I image (Figure 3

Component 2 and Its Loading Scores
This component image and its graph of loading scores ( Figure 5

Component 4 and Its Loading Scores
This is the third vegetation change component which shows about 2% of the total variance in the total time-series dataset ( Table 1) October of 1999 ( Figure 6(b)). On the other hand, the sandy soils in most of the arid and semi arid areas in the study area are usually low in organic matter, nitrogen and phosphorus also degrade rapidly under conditions of intensive rainfall [34]. Furthermore, based on the findings by [35] and, because of the nature of the rainfall in the front-line states which support mostly savanna vegetation the density of trees and other plants tend to be decreasing as one moves northwards. This is another reason why the savanna ecosystem of this zone in the country is very sensitive to human and animal population pressure. As [36] re-

Conclusion
From the results derived from this broad assessment, it can be concluded that dataset acquired from SPOT satellite for the purpose of monitoring the status of vegetation as a resource can be very useful for consistent and periodic monitoring of land cover generally. In particular, issues relating to climatic impacts globally, continentally, regionally, nationally and even at local levels can be undertaken. At the national level for example, a broad assessment using time-series of medium spatial resolution dataset can be undertaken. Once a systematic methodology is devised, adopted and maintained it can go along way for the policy makers particularly in monitoring vegetation as a resource. Areas of likely hazards such as drought and floods as well where serious disasters occurred can be analysed for sustainable environmental and resource management. Further research should include socio-economic data and extensive ground-truthing where area covering Nigeria should exclusively be analysed.