Faidherbia albida (Delile) Tree Dieback Effects on Crop Production in the Parkland Agroforests of Southwestern Niger

Faidherbia albida is an agroforestry tree species playing important agroecological and socioeconomic roles in arid and semiarid zones in Africa. For many years, anthropogenic and abiotic stresses were considered as the main threats for the species in West African parkland agroforests. Considerable dieback has recently occurred in F. albida trees of parkland agroforests in central southwestern Niger, and the causes are unknown. The objectives of this study are to (i) investigate the magnitude of dieback of F. albida trees and (ii) assess local community perceptions of the eﬀects of F. albida dieback on crop production. The health status and phenology of 213 F. albida trees were observed in the area where the dieback is occurring. Similarly, a sample of 144 people, 86% of which were farmers, was surveyed. Dieback incidence of F. albida trees was 19%, with mortality of 6%. Large-diameter trees had greater dieback than small-diameter trees. The most aﬀected parts of the tree were the branches at 54% and the trunks at 39%. The populations noted a 33–55% reduction in the yields of major crops. This dieback of F. albida trees poses a serious threat to the survival of rural communities. Further studies can be conducted to identify the cause or cause of the dieback to guide the suitable agroforestry parkland management strategies.


Introduction
F. albida, originally known as Acacia albida as named by Delile [1], belongs to the family of Fabaceae [2]. Under favourable conditions, F. albida can reach more than 30 m in height and 1.5 m in diameter at breast height. e crown is pyramidal when the tree is young and becomes hemispherical with age. e leaves are compound and bipinnate. F. albida is a widely distributed species in the arid and semiarid zones of Africa [3,4]. Unlike most Sahelian woody species, F. albida is a reverse phenology species: it loses its leaves with the first rains of the season and regrows them during the dry season [5,6]. On the basis of seed analysis, Ibrahim et al. [7] highlighted three ecotypes of F. albida in Africa: the Sahelian ecotype, the South African ecotype, and the ecotype that overlaps between the two zones, particularly in Ethiopia. In the same ecological zone, F. albida has contrasting morphological and growth variability [8][9][10][11][12]. e trees have long been used traditionally in the principal agricultural production system in West Africa, known as parkland agroforest. F. albida is an excellent agroforestry species for its impact on improving soil carbon and soil fertility, increasing the activity of symbiotic microorganisms, and improving the yields and nutritional values of crops [3,[13][14][15][16][17]. Due to its deep root system in dune soils, the competition for superficial water and nutrient resources between this species and surrounding crops is very limited [11]. e F. albida parkland agroforests of central southwestern Niger were born from the intervention of forestry projects promoting the species beginning in 1981. Because of the numerous benefits of F. albida trees mentioned above, the projects have the common goal of increasing their density from 35 to 50 trees/ha based on the low-cost investments of farmer-managed natural regeneration (FMNR) and planting wherever possible. Conclusive results have been obtained, leading to the domination of the parklands by F. albida. For example, the yield of millet grown under foliage of F. albida trees was 1.78 times higher than the yield of comparable plots without these trees [18]. During the last several years, however, this species, which improves the well-being of these local communities, has experienced increasing pressure not linked to human exploitation, but to a dieback leading in most cases to the trees' mortality. Similar attacks leading to the mortality or dieback of both young and old F. albida trees in the 1990s were reported by [18].
Tree dieback can be defined as the expression of the immediate effects of acute stresses [19]. e causes of this dieback remain unknown. However, many studies have addressed the decline of trees in plantations, natural forests, and farms around the world. e causes are still debatable. Very often the most indexed causes have been folivores, infections, and abiotic stresses which can be linked mainly to climate or soils [20,21]. Rural tree dieback could be linked by factors such as global climate change, overuse of pesticides, heavy metals in soils, overuse of fertilizers, pathogens, and habitat fragmentation [22,23].
ese factors have direct effects on the survival of symbiotic microorganisms (mycorrhizae) which play an important role in the life of trees. Based on all of these causes, theories and models of tree dieback have been developed [24][25][26]. e interaction between several factors including biotic, abiotic, and human makes the study of tree dieback very complex. us, in his spiral of tree dieback and mortality, modified by [27,28], main groups of causes are presented as follows: (i) predisposition factors, namely, the tree genetics, the soil nutrient deficit, climatic variability, chemical, physical, and biological degradation of soils, and chronic attacks from pathogens and insects; (ii) incentive and contributing factors such as extreme climate, high soil toxicity, insect defoliation, and competition. e dieback of F. albida trees was most often observed in the Dogondoutchi department in Niger, especially in the villages of Kiéché Commune. ere is an urgent need to understand the causes and implications of the dieback so that effective responses can be put in place. Long before, it would be necessary to understand the magnitude of this tree dieback in this complex of agroforest parklands. To contribute to this, we asked the following questions: what is the magnitude of dieback of F. albida trees in these parkland agroforests? What is the current perception by local communities of the effects of this dieback on the main crop yields in the area?

Sites.
e study area is located at 13°28′ N and 4°3′ E with an elevation of 311 m [29], on the left bank of the Dallol Maouri, a broad fossil valley framed by sandstone and sometimes cuirassed plateaus [30]. It is an agricultural area par excellence. e climatic data provided by the national meteorology department of Niger were used to characterize the climate of the study area. e climate is of the Sahelian type, with an average annual rainfall of 540 ± 98 mm and maximum and minimum average annual temperatures of 36.8°C and 23.2°C, respectively. e rainy season typically lasts 3-4 months (June-September). e maximum and minimum average annual humidity are 59% and 34%, respectively.
e study area belongs to the continental terminal formation with a spatial and temporal variability of the phreatic aquifer. Piezometric fluctuations throughout the year range from 1 m to 9 m [31]. e vegetation is dominated mainly by parkland agroforests long enriched by plantations of F. albida trees and the practice of FMNR [18]. e study sites were chosen following interviews with the officers of the Department of the Environment in Dogondoutchi and the Rural Municipality of Keché. is choice was also based on the presence of F. albida trees and the prevalence of dieback in the parkland agroforests. is information made it possible to compile a list of five villages (KalloMota, Ruda, Adoua, GarinPaydou, Tambo Gataw, and GarinDianta) suitable for the study and distributed in different geographical locations of the municipality.
Two methods were used: surveys of local communities and an inventory of the parkland agroforests of the selected sites.

Surveys.
e surveys were conducted in two stages in July 2018. Group discussions were conducted in each village to capture the overall perception of the problem, and individual surveys were conducted to obtain quantitative data using random sampling. e group discussions were focused on (i) the importance and benefit of F. albida trees for local communities; (ii) the threats to F. albida trees; (iii) the causes of the mortality observed in F. albida trees; and (iv) the manifestations of tree dieback. Individual surveys took a random sample of 144 people from all villages of which 86% were farmers. e 14% are composed of pastoralists, traditional healers, blacksmiths, etc. due to the exploitation of F. albida for their various uses. e average age of those surveyed varied from 47 to 50 years, and all socioprofessional groups (level of education, primary and secondary activities, etc.) in each village were represented. e surveys also examined the effects of F. albida tree dieback on the yield of the main crops in the area. e questions asked included farmers' estimates of crop yields before and after the mortality of F. albida trees in the different farms of the parkland agroforests. e main crops were Pennisetum glaucum (millet), Sorghum bicolor (sorghum), Vigna unguiculata (cowpea), and Hibiscus sabdariffa (sorrel of Guinea). e yields of each crop were given in the local unit and then converted to the international unit (kg/ha). It was difficult for the respondents to evaluate the crop yield specifically in the area under the tree crown before and after the dieback. us, the evolution of the yield was reported at the farm scale of each respondent.

Inventory.
e tree inventory was conducted in July 2018 to provide data on the distribution of woody species, including F. albida trees, the trees' current state of health, and the prevalence of dieback. e parkland agroforests of three of the five villages were inventoried. e two others were close to each other and therefore provided no variability. Radial sampling with transects from the village in the four geographical directions was used [32,33]. A total of 48 plots were installed with 16 plots per village. Square plots measuring 50 m × 50 m were installed [34]. e distance between two plots was 300 m. Within each plot, an exhaustive count of woody species was performed, including F. albida trees. For each F. albida tree, observations were made on the health status (healthy, sick, or dead) and phenology (leafing, flowering, or fruiting). It would be difficult by simple observation to dissociate diseased trees from healthy trees. For dead trees, several definitions have been reported in the literature [35]. However, for this study, we are retaining dead trees, including fallen or standing trees, when its aerial part (leaves, branches, and trunk) has dried out or at the limit is rotten while trees are sick or stressed when they suffer damage from biotic and/or abiotic agents [36] and show signs of drying out or attacks on at least one of its aerial parts [37]. us, observations were made on all sampled trees, and medium diameter trees were the most attacked ( Figure 1). Samples of leaves, branches, and roots of attacked trees and soil near the roots were taken for laboratory analysis to detect possible pathogens. e results are not presented because they are not the subject of this article.

Data Analysis.
e survey data were processed using an Excel spreadsheet and Minitab 16 software for descriptive statistics (means, standard deviations, and frequencies). e one-way ANOVA test was applied to determine the statistical significance of the evolution of crop yields according to the periods before and after the mortality of F. albida trees at 5% of significance.
is test was used to address the heterogeneity and nonnormality constraints of the data and sample size [38]. Finally, the yield difference according to the two periods was reported in percent.

Local Perceptions of the Importance of F. albida Trees.
Farmers were unanimous that the presence of F. albida trees improves soils fertility, protects crops against wind erosion, and controls the movement of sand. e species is also an important source of animal feed; farmers estimated one F. albida tree produces, depending on its age, an annual average of 90 kg of fruit pods without succulent branches at a price of 14,565 ± 2,383 FCFA (equivalent to 24.27 U. S. dollars).

Magnitude of F. albida Tree Dieback.
Observations on the health status of F. albida trees in parkland agroforest yielded 74% of healthy trees versus 19% sick and 6% dead but standing. About six F. albida trees per hectare were sick or dead but standing (Table 1; Figure 2).

Local Perceptions of Effects of F. albida Tree Dieback on
Crop Yields. Populations have recognised the loss of tree populations over the past decade, and this issue is becoming increasingly important in the area. e dieback concerns only F. albida trees; this is also a function of age. According to the respondents, the yields of the main crops mentioned were affected by the dieback of F. albida trees in the area. In fact, the average yields in kg/ha before the dieback were 386.21 ± 8.92, 634.65 ± 34.90, 161.11 ± 11.64, and 90.91 ± 20.42 for millet, cowpea, sorrel of Guinea, and sorghum, respectively. Following the dieback, yields experienced a drastic reduction ranging from 33.03% to 55.47% (Table 2).

Discussion
e parkland agroforests of F. albida in Dogondoutchi were born out of the will to face the major challenges of environmental protection and the improvement of agricultural production. e last one remains a challenge in Niger due to significant population growth (3.6%), food, and nutritional insecurity, the drop-in yield as a result of the decline in soil fertility, and parasitic pressures. Developing agroforest parklands where trees provide soil protection and fertilization remain necessary for the sustainability of the system. Significant investments have been made through development projects since the 1980s to increase the density of F. albida trees from 35 to 50 trees per hectare in the area [18]. e local populations agree that F. albida trees have significant positive effects on soil fertility, fodder for animals, and crop yields. A meta-analysis showed significant improvements in soil organic carbon (46% increase), total nitrogen (50% increase), phosphorus (21% increase), potassium (32% increase), and yields of maize and sorghum (150% and 73%, respectively) under forest cover relative to the dryland zone [16]. In Niger, [14] reported an improvement in millet yield of 36% under F. albida trees. is is consistent with farmers' perceptions in this study: they observed a fall in their crop yields of between 33% and 55%, depending on the crop, with the mortality and the disappearance of F. albida trees.
In general, tree dieback has been addressed in many studies and causal links have been discussed [24,25,26,39]. In the present study, F. albida trees most affected by dieback are those with large and medium diameters at breast height and with a density of less than 6 trees/ha. Does this mean that the dieback would be linked to the age of the tree? is statement is difficult to accept, but not to exclude, because in F. albida, the size of the tree is not necessarily related to age [40]. Physiologically, this could be explained by the fact that, in arid areas, when woody species are stressed, a trade-off emerges between safety and efficiency. e tree tends to favour its survival functions to the detriment of that of growth or reproduction by eliminating leaves, small twigs, and at worst that of large twigs and trunk [37,40]. e most affected parts of the tree were branches at 54% and trunks at 39% of cases. One of the direct consequences of this dieback on local communities was a reduction in the provision of ecosystem services by these parkland agroforests. e most striking aspect was the yield reduction in the area, which ranged from 33% to 55%, depending on the crop (Table 2).
is decline was attributable to the reduction in the nutrient cycling that F. albida trees provide.

International Journal of Forestry Research
According to the perceptions mentioned by the local communities in our study area, there are three main causes of mortality of F. albida: defoliating caterpillars, the larvae of beetles that were found in the trunks of dead trees, and rising groundwater. ese causes are easily compared with those evoked in the spiral of [26,38] which gives a general theoretical framework where all the plausible causes have been evoked with complex links leading to dieback and death of trees. is complexity lies in the interaction between biotic (bacteria, fungi, viruses, etc.), abiotic (climatic variability, soils, water stress, etc.), and human factors (modes of landscape occupations, tree management, etc.) [20,21]. e causes mentioned by the populations related exclusively to pest and disease part factors of the spiral. Analyzing these causes given by the populations shows them to be merely visual symptoms of the causes, but positive or negative interactions are created among the different living beings (e.g., microorganisms and trees) that are at times invisible to the naked eye. Symbiotic interactions favourable to tree species including F. albida canopies have been demonstrated [13]. On the other hand, those leading to the mortality of tree species have also been reported. For example, Fusarium solani has been reported in Acacia nilotica in Pakistan [41]. Similarly, a fungus named Pseudolagarobasidiumacaciicola has been identified as the causal agent of the 100% Acacia cyclops stand mortality in South Africa [42]. On the other hand, studies carried out in Namibia showed a mortality of F. albida trees of 51% of the stand, which was linked to the variation in the flow of water and the presence of invasive Prosopis species [43]. e first cause corroborates the statements of the local populations of our study area, while the second does not fit because F. albida is the dominant species and coexists with native tree species (not exotic, invasive Prosopis species) at the three sites. On the other hand, the authors in [4] have reported the presence of nematodes and defoliating caterpillars causing the mortality, especially of F. albida trees. Ultimately, all avenues must be explored to better detect the disease and/or pest of this species in the study area.    Hence, there is a need to take seriously the problem of dieback to avoid its expansion towards other ecological zones, because F. albida is an important agroforestry species with a wide distribution in the dryland zones of Africa [44].

Conclusions
is study has shown that the parkland agroforests of Dogondoutchi, dominated by F. albida trees, were affected by the dieback. e F. albida trees were wasting away, with an infestation rate of 19% and a mortality rate of 6%. e causes of this dieback are scientifically unknown. According to local communities, this dieback reduced the yields of their main crops by between 33% and 55%. To explore also the causes of the dieback of F. albida trees, it is important to take into account the causal links presented by [27].
is involves analyzing samples taken from parts of infested trees to isolate the presence of biotic factors, soil samples for the plausible presence of heavy metals, or nematodes.
is study provides field information on the current dynamics in the parkland agroforests of central southwest Niger. is information can be used (i) for famers to make efforts to ensure the regeneration of this species in their parkland agroforest and (ii) for scientists to conduct in-depth studies to identify the real causes of mortality of F. albida, whose large trees are regularly attacked, to test and identify plausible solutions in collaboration with local communities.

Data Availability
e data used to support the findings of this study cannot be shared because they are confidential.

Conflicts of Interest
e authors declare that they have no conflicts of interest. International Journal of Forestry Research 5