The impacts of anthropogenic fires in West African savanna woodlands and parklands: the case of the Guinea savanna, Ghana

Fire is recognised as an important factor influencing the structure and function of tropical savannas. Despite the extensive studies conducted on the effects of fire on global savannas, there are relatively few studies focusing on fire-assisted land use practices in the agroforestry parkland of the Sudano-Guinean savannas of West Africa. The region experiences recurrent fires in the dry season which begins from November to April. The fires are anthropogenic and are mainly caused through rural livelihoods and cultural practices such as farming - to remove debris from crop fields and to improve soil fertility and hunting to flush out animals, among other reasons. This study therefore sought to: 1. contribute to the understanding of fire-assisted traditional land use practices, people’s knowledge and perceptions of fire use and fire regime: 2. analyse the effects of fire on vegetation, and 3. determine the effects of fire on soils in the Guinea savanna woodlands and parklands of Ghana. Study sites (burnt and unburnt land use types) were selected based on five-year daily fire counts (2013-2017) data, obtained from the Earth Observation Research and Innovation Centre (EORIC), Ghana in collaboration with the Council for Scientific and Industrial Research (CSIR) Meraka Institute, South Africa. Fire densities were calculated for the 18 districts in the Guinea savanna (Northern Region of Ghana). The districts were stratified into low, medium and high fire frequency areas. The East Gonja district recorded the highest fire density (1.0 fires km-2) while Tamale recorded the lowest fire density (0.3 fires km-2). Of the eighteen districts, six districts were purposively selected and ten communities sampled. Firstly, I investigated the frequency of fire use and control, perceptions of fire regime for selected livelihood and socio-cultural activities in the six districts. The majority of respondents (83%) across the study districts indicated that they used fire once a year for at least one of the following activities: land preparation, weed/grass/pest control, burning stubble after harvest, bush clearing around homesteads, firebreaks, charcoal burning and hunting. The study showed the highest frequency of fire use in the dry season was for land preparation for cropping. However, less than a fifth of the respondents (17%) indicated that they did not use fire for any of the above activities. The results of a multiple regression predicted fire activities in the dry season from gender, age, level of education, occupation and household size. Secondly, the study examined how fire influenced the population structure and abundance of two economically important woody species Vitellaria paradoxa C. F. Gaertn. (Shea tree) and Anogeissus leiocarpa (DC.) Guill. & Perr. (African Birch) in burnt and unburnt land-use types. Stand basal area, mean densities of juveniles and adult trees and Simpson’s index of dominance were determined. Eight diameter size classes of each species were analysed by comparing their observed distributions to a three-parameter Weibull distribution across the land use types. A total of 3,366 individuals of A. leiocarpa (n = 1,846) and V. paradoxa (n = 1,520) were enumerated. The highest total basal area of Anogeissus leiocarpa (16.9 m2/ha) was estimated in sacred groves whereas Vitellaria. paradoxa (20.6 m2 /ha) was found in unburnt woodland. The highest mean densities of A. leiocarpa (22.7±29.7 stems/ha) and V. paradoxa (15.3±2.2 stems/ha) were found in sacred in groves. Anogeissus leiocarpa was, however, absent in fallows and burnt crop fields. A somewhat inverse J-shaped distribution was found in sacred groves for both species. Thirdly, the study compared species composition in early burnt, late burnt and unburnt plots in a protected area where fire was regulated. Vitellaria paradoxa C. F. Gaertn., Terminalia avicennioides Guill. & Perr., Combretum adenogonium Steud. ex A. Rich. and Combretum molle R. Br. Ex. G. Don. were the most common and abundant in all treatments. Late burnt and unburnt plots recorded the lowest diversity amongst the three treatments. Unburnt plots had higher tree density than burnt plots. A Detrended Canonical Analysis showed a changing trend, indicating a moderately strong positive association between burning time and species composition. The first and second axes contributed 53% and 12% variation, respectively. Most of the species found in axis one had fairly strong positive association to early and late burnt treatment than unburnt treatments. Finally, the effects of fire on soil properties in burnt and unburnt crop fields and woodland in the ten communities were also investigated. A total of 151 composite samples in selected burnt and unburnt land use types (burnt crop field n=20; unburnt crop field n=27; burnt woodland n=53, unburnt woodland n=51) were analysed. The variables analysed were pH, soil organic carbon (SOC), total nitrogen (TN), available phosphorous (P), Exchangeable bases - potassium (K), magnesium (Mg), calcium (Ca), cation exchange and soil texture. Total N, SOC, pH and Ca differed significantly across the burnt and unburnt land-use types. A principal component analysis showed a stronger association and more positive gradient in woodlands than in crop fields. Total N showed a positive association with SOC, whereas silt showed a negative association to sand and clay. Traditional knowledge and perceptions of users of savanna agroforestry parklands can inform the formulation of local by-laws for community fire management as well as national policy regulation on fire use in the savanna through the interlinked analysis of social and ecological systems as have been elucidated in this study. The results on vegetation assessment revealed that fire practices and land uses influenced size class distribution of the two study species as well as the densities of woody species in traditional crop fields and the National Park. The unstable populations observed in most land uses and the absence of A. leiocarpa in crop fields and fallows call for education and policy actions on the use of fire in parklands of West Africa. Furthermore, fires positively influenced some soil properties in both woodlands and crop fields confirming one of the emphasised reasons why rural subsistence farmers use fire. As an environmental management decision and land use policy intervention, early dry season burning could be an option to curbing the indiscriminate and unplanned fire use. Also, the protection of trees on farm lands could check the unstable population structure of economically important woody species leading to the different structures that deviated from the recommended reverse J-shaped distribution curve observed in these agroforestry parklands. The management of socio-ecological systems such as the agroforestry parklands of West Africa require a holistic understanding of the complexity of the different resource systems, units and actors involved for sustainable management of these natural resources.