2.1. Study Area

The study was conducted in Lilongwe district, which harbors the capital city of Malawi, located in the central region of Malawi between latitude 13.59 south and longitude 33.47 east (Figure 1). Its elevation ranges from 1000 to 1200 m above sea level, with a primarily flat terrain (GoM, 2010). The city has four major sectors, namely, Lumbadzi, Capital Hill, Kanengo, and Old Town [10].

The city’s climate is characterized by two distinctive seasons: hot–wet season from November to March, and the cool–dry season from April to October. The city experiences an average annual rainfall between 800 and 1000 mm and a mean annual temperature of about 20°C–22.5°C . The district covers an area of 615,900 ha, of which, 37,786 ha is delineated as the main city. In terms of geology, the region is mostly composed of metamorphic and plutonic rocks, which are topped by alluvium and soil made of worn metamorphic and plutonic rocks [8].

In 2005, Lilongwe City experienced a substantial increase in urban growth, which was further enhanced by the repositioning of government offices from Blantyre City to Lilongwe. The current city’s population is estimated at 1.1 million, with an estimated annual growth rate of 4% [12]. With regard to social infrastructure and fundamental urban services, nearly 76% of settlement areas are deemed unsafe. This presents a significant threat, leaving residents vulnerable to natural calamities such as floods and fires [8].

In addition, the city acts as the central region’s main hub for agricultural products’ markets. The primary industrial region is Kanengo in the north, which is the main center for light industry operations such as food processing, maize storage, and tobacco sales and storage, among many other industrial activities. Parks and recreation centers also contribute significantly to the city’s economy. Even so, the unemployment rate is estimated at 16% leading to approximately 25% of the inhabitants being poor [8].

2.2. Data Collection Methods

A stratified random sampling method was employed in order to obtain the required samples. This sampling procedure ensures that homogeneous units, which capture key subgroups and minority groups of tree communities, are created and hence improve the precision of the study [13]. The strata were determined according to the existing informal urban land use types specifically where urban forests fall [8, 9]. According to informal categorization, urban forest types were categorized based on the purpose of forest types as outlined in Table 1 [8].

In brief, in residential areas, random sampling was performed by considering population sizes. Residential areas were categorized as low-, medium-, and high-density areas. Two areas were selected in both low and medium areas, while three areas were selected in the high-density area. In each area, four houses were purposively sampled from the houses with a considerable number of trees [9]. A complete tree inventory was performed accordingly at each sampled household.

For urban street trees, roads were sampled as determined through Google Earth Map and verified by the LCC authorities or the Malawi Roads Fund Administration (MRFA). First, random sampling was performed to select the required number of streets. Eleven avenues were identified. The first two were the major streets of ≥ 1000 m across the city and hence were automatically included. Then, three avenues were purposively selected from the remaining nine. In each avenue/road, sample plots were laid within a stretch of 20 m long on both sides of the road/street at a regular interval of 500 m.

For riverine/riverbank forests, Lilongwe City has two major rivers that pass through it [8]. Hence, two major rivers were identified and purposively selected as Lilongwe and Namanthanga Rivers, respectively. A distance of 20 m (riparian zone) for each selected river was sampled. Along each riverbank, sample plots were laid within a stretch of 20 m long on both sides of the river at a regular interval of 500 m within the city.

Lastly, four institutional areas were randomly selected with a subcategorical consideration of each for offices, schools, churches, and lodges. No sample plots were laid in the identified areas, rather the area for each institution was measured, and a complete tree inventory was performed. Table 2 and Figure 2 summarize the sampled areas.

In each sample plot or sampled area, GPS coordinates were taken and a full inventory was conducted for all tree species as required. According to FAO [4], a tree is well-defined as any perennial woody plant possessing one main trunk, or in the case of coppice, with numerous trunks, possessing a more or less definite crown. In this case and for the purpose of this study, a tree was regarded as such excluding bamboos, palms, and other woody plants meeting the criteria in question. Thus, in each sampling plot, an inventory of trees taller than 1.5 m and of diameter at breast height (DBH) (1.3 m from the ground) of greater than or equal to 5 cm was performed. Trees in each sampling plot were identified up to species level with the help of a plant taxonomist and guidelines of plants of Malawi and nearby countries [16] as well as the Malawi species code list developed by the Forest Research Institute of Malawi (FRIM) and the National Herbarium. The sample of all unidentified tree species specimens was collected, pressed, dried, mounted, and deposited in the Herbarium of FRIM with a code number on the description sheet. Furthermore, tree identification relies on taxonomic classification, which organizes trees into hierarchical categories: kingdom, phylum (division), class, order, family, genus, and species. This system enables accurate identification and comparison by grouping trees based on shared characteristics, aiding botanists, ecologists, and foresters in research, conservation, and sustainable management practices [9].

2.3. Data Analysis

Tree species richness and diversity were assessed using Rẻnyi diversity profiles in BiodiversityR. [17]. The functionality of BiodiversityR has been thoroughly explained by Missanjo et al. [18]. In summary, BiodiversityR is a software that performs comprehensive biodiversity analyses, while Rẻnyi diversity profiles are curves that provide insights into species richness and evenness. The shape of the profile indicates evenness, with the initial position on the left-hand side representing species richness. A profile that starts at a higher level indicates greater richness. The main advantage of Rẻnyi diversity profiles is that they allow for easy comparison of sites in terms of diversity. If the profile for one site is consistently above that of another, the site with the higher profile is more diverse.

3.1. Tree Species Richness in Urban Forest Types in Lilongwe City

The Rẻnyi diversity profiles for the six forest types in Lilongwe City are presented in Figure 3 and a summary of tree species encountered in each forest type is presented in Table 3.

A total of 498 sampled plots were laid in all the six forest types constituting a total record of 4031 individual trees encountered (Table 3). The number of tree species was 165 which were recorded in all six forest types (Table A1). Interestingly, of the total tree species recorded, only five tree species (Gmelina arborea, Khaya anthotheca, Rauvolfia caffra, Senegalia polyacantha, and Senna spectabilis) were recorded in the six forest types. Each of the six urban forest types had its own tree species, which are not shared with any of the forest types. In this respect, 38 tree species were recorded in parks/recreation forests, 37 tree species were recorded in residential forests, 15 tree species were recorded in institutional forests, 4 tree species (Vachellia karroo, Ficus exasperate, Ficus natalensi, and Tecomaria capensis) were recorded in road/avenue forests, while one tree species (Afzelia quanzensis) was recorded in the cemetery forests.

Furthermore, the most dominant family consisting of at least four species was identified (Table 4). These families constituted 64.44% of the total tree species documented in the six urban forest types in Lilongwe City.

The parks and recreation and residential area forests had the highest species number probably due to the protection they receive and that most of these forests lie in Miombo woodland forest areas [19] in addition to being supplemented by planting of exotic tree species [9]. Less number of tree species was recorded in institutional and road/avenue forests. This could be a result of the loss of most trees due to construction as the land is cleared and the planted trees are mostly chosen for a specific purpose, which is mostly ornamental. This also concurs with what the authors in [9, 20, 21] concluded that there is a great anthropogenic influence in these land use types, hence a smaller number of tree species as the choice is in accordance with the owner’s interests.

The similarity of most tree species being found across all the forest types is attributed to the indigenous forest type of area being Miombo woodland [8]. Most species are of indigenous type as they have been naturally established [19], whereas, Afzelia quanzensis, which is found only in cemeteries, could be a true reflection of this species being sacred in Malawian culture.

3.2. Tree Species Diversity in Urban Forest Types in Lilongwe City

The results on tree species diversity (from Rẻnyi diversity profiles value), which were analyzed based on richness and evenness levels in each forest type, showed that the residential forest had a higher Hα (4.47) at 0-alpha followed by the parks/recreation forest (Hα = 4.41) (Figure 3). The riverine forest had the lowest Hα (2.71) at 0-alpha. This indicates that the residential forest had a higher tree species richness and evenness, based on Rẻnyi diversity profiles, followed by the parks/recreation forest, while the riverine forest had the least tree species’ Rẻnyi diversity index value (richness and evenness). Households have proved to harbor a considerable number of trees in accordance with similar studies conducted in Zomba City, Malawi [9], and Kumasi City, Ghana [22].

The results show that the Rẻnyi diversity profile for residential forests is higher than those of the other five forest types (Figure 3). This is an indication that residential forests are more diverse than the other forest types. However, the shape of the profile for the residential forest is less horizontal. This is an indication that the proportions of the individual tree species are not evenly distributed. This was attributed to the dominance of one species (such as Mangifera indica). According to studies [9, 22] in Malawi and Ghana, respectively, residential areas indeed have a higher number of species. However, they are less diverse because the choices of the tree species are mostly determined by the owners. It has been observed that most tree species in households are those which can provide food/fruits such as Mangifera indica. This fruit tree was found to be dominant in almost all the households of the residential forest type.

According to other studies [9], Mangifera indica, Psidium guajava, Persia americana, and Carica papaya were the dominant species in households, signifying their purposes for food beyond their values as beautification of the homesteads. Similarly, the study [23] concluded that tree households’ spaces in most cities of developing countries were planted for a variety of purposes, including fruits, nuts, fuel wood, vegetables, fodder, shade, and windbreak. For this reason, the authors in [24] argue that trees and their seed networks have implications for genetic diversity with regard to a study on urban fruit trees conducted in Yaounde, Cameroon.

Similarly, the shapes of the diversity profiles for the other five forest types are less horizontal. This means that the proportional of the individual tree species were not relatively evenly distributed [25]. An even distribution of individual plant species in an ecosystem signifies a healthier or good ecosystem where few plant species are not dominating the community/vegetation [26]. As stated by the authors in [25], a denser, functionally diverse, and well-adapted community of urban trees could provide more ecosystem services that are more resilient to the increasing risks of biophysical disturbances induced by environmental changes. Even with the limitations of available space for infrastructure and buildings, there are still plenty of opportunities for planting trees, even though forest-like tree densities are not appropriate for all land use types. For instance, the authors in [25] suggest that through hybrids of traditional and green infrastructure designs, it may be possible to increase tree density and diversity by using soils that are good for load bearing and root development in sidewalks. McKinney [20] concurs that, such alterations to street tree growth environments may make it easier to plant a wider variety of species, including more indigenous tree species; in essence, these kinds of indigenous tree species preservation and promotion initiatives can prevent the loss of regional biotic distinctiveness. On the other hand, modifications to landscaping techniques can present chances for spontaneous regrowth [27].

Therefore, the present results suggest that all the six forest types are disturbed and need appropriate management. This is in line with the findings of a similar study carried out in Kumasi, Ghana, by the authors in [21], which showed that changes in land use and land cover between 2007 and 2017, which are primarily related to infrastructure development in most cities, have significantly decreased the richness and diversity of urban tree species. However, the major challenge remains in the minimal understanding of drivers of diversity about different types of urban forests that make up the entire urban forest ecosystem, which represent differences in tree planting and establishing techniques, ownership, and maintenance [28–32]. For instance, in a study conducted, by the authors in [2], in cities of Lilongwe and Mzuzu, Malawi, it was found that there is negligible evidence that urban green spaces and green setup are integrated into decision-making protocols at the national level. The planning and strategy documents created by the city councils of Mzuzu and Lilongwe did, however, prioritize developing and expanding urban green space and green infrastructure. Therefore, the analysis suggests that if the numerous advantages of urban green spaces and green infrastructure are to be realized, there has to be better institutional coordination and policy coherence across national level sectors that are impacted [2].