The role of infrastructure and route type choices for walking and cycling in some cities in Ghana

ABSTRACT Infrastructure has contributed to eradicating poverty and improving the livelihood of people. In Sub-Saharan Africa, active transport infrastructure is less, poor, or lacking, which has become one source of the decline in cycling and efficiency in walking. This study sought to investigate the types of infrastructure or routes that active transport users are exposed to, determine their usage magnitude and solicit views regarding them through questionnaire surveys and screen-line counts in two major cities in Ghana. Active transport respondents who comprised of pedestrians and a few cyclists identified themselves with these infrastructures and route use frequently: streets, sidewalks, cycle lanes, unpaved footpaths, and ‘lungulungus’. Street and ‘lungulungus’ were the most used infrastructure or route choices by all age, gender, and occupation groups in both study locations at Ayawaso East Municipal Assembly and Tamale Metropolitan Assembly. The respondents chose to use these two mainly because they were direct, cohesive and safe. The fact that people used ‘lungulungus’ more than the provided infrastructures suggests poor planning of the built environment and inadequate dedicated infrastructure. This calls for a review of existing dedicated infrastructure and the provision of infrastructure that addresses the needs of pedestrians and cyclists.


Introduction
The built environment is the pivot for active transport growth and patronage and there is, therefore, the need to create an environment that protects and promotes walking and cycling, which are known forms of clean and cheap transport (Marshall & Garrick, 2010;Mulvaney et al., 2015;Tiwari & Jain, 2013).Walking and cycling are also sustainable modes with accompanying health and social benefits (Winters et al., 2011).Urban forms in higher-income countries over decades were built to suit auto-oriented development that has not supported safe road traffic environments, especially for non-motorised transport users (Schepers et al., 2018).In the United States, cycling is reported to be more dangerous than car travel (Winters et al., 2011).The rate of fatality per bike trip was about 2.3 times higher than that of automobile trips due to the unsafe environment (Winters et al. 2018).The same has been reported in the United Kingdom, where walking and cycling are said to be twice as dangerous as in Germany and Denmark (Buehler & Pucher, 2012).Between 2005 and 2014, of the 253 cyclist casualties in the three cities of northern Ghana (i.e.Tamale, Wa and Bolgatanga), Tamale accounted for 26% of which more than half (55%) died (Damsere Derry & Bawa, 2017).According to Acheampong and Siiba (2018), the safety risks that were deterrents to cyclists or non-cyclists were related to the physical environment, which includes infrastructure and traffic conditions.Notwithstanding the benefits of walking and cycling, these sustainable transport systems are generally on the decline globally (Loo & Siiba, 2019), and safety is one of the reasons for the decline or abandonment (Quarshie, 2004) Safety issues mostly confronting cyclists and other non-motorised transport, according to Kingham et al. (2011), are primarily due to the lack of compatibility between motorised traffic and non-motorised traffic infrastructure concerned.Infrastructure is one of the major environmental features that creates opportunities for walking and cycling and further growth in cycling and improvement in walking (Alliance for Biking & Walking, 2014;DiGioia et al., 2017;Jennings, 2015;Tiwari et al., 2016;Vale et al., 2016).Infrastructure is one of the performance indicators for sustainable transportation (Litman, 2017).Zhou et al. (2020) acknowledged that the provision of a good nonmotorised transport (NMT) facility is an effective way to promote bicycle use and subsequently improve one's physical health.In countries like the Netherlands, Denmark and Germany, the provision, expansion, and improvement of good cycling and walking facilities as a public policy have played a role in the support and growth witnessed in walking and cycling (Buehler & Pucher, 2012;Pucher & Buehler, 2008;Pucher & Dijkstra, 2003).The extensive bikeway network, coupled with the hospitable environment, is said to have encouraged cycling in Bogota (Cervero et al., 2009).In Europe, America and other developed countries, surges witnessed in cycling and walking can be attributed to a lot of measures including the provision, increment or improvement of infrastructure (Buehler & Pucher, 2012;Cobbinah et al., 2018;Nkurunziza et al., 2012;Ribeiro et al., 2007;Srinivasan et al., 2019).In developing countries, active transport, especially walking, is popular (Amoako et al., 2014), and cycling is prevalent in rural areas (UNEP, 2016).However, the infrastructure for the two modes is limited (Jennings, 2011;Mogaji & Uzondu, 2022;Oyeyemi et al., 2013;Mokitimi & Vanderschuren, 2017).Loo and Siiba (2019) attribute the low cycling levels in African countries to the lack of cycling infrastructure.Good infrastructure provision has therefore come up among other recommendations as one of the solutions to promote and curb the safety risks associated with cycling and walking in Africa (Acheampong & Siiba, 2018;Damsere Derry & Bawa, 2017).The notion emphasizes the relevance of infrastructure as one of the important requirements for active transport in sub-Saharan Africa.
The active transport infrastructures are either separated or attached to motorised streets and come in names like cycle lanes, sidewalks, bikeways, and pedestrian walkways (CROW, 2016;Loukaitou-Sideris & Renia, 2011;Mateu & Sanz, 2021;NACTO, 2011;Yannis & Chaziris, 2022) with a physical separator or delineators (Mateu & Sanz, 2021).They are either exclusively used by pedestrians or cyclists or in a mixed form.Infrastructure for walking and cycling usually has no identified safety problems except for being inefficient depending on the capacity of the infrastructure and volume of the active transport users where they exist as a mixed type for both pedestrians and cyclists (I-ce, 2000).There is a positive correlation between infrastructure and active transport (Adams et al., 2013;Hull & O'Holleran, 2014;Pucher & Buehler, 2006), making the presence of infrastructure important for the growth of active transport.

Active transport infrastructure challenges
The condition of the pedestrian or cycling infrastructure is as important as its availability regarding the choice to use active transport modes.According to standards such as NACTO's (2011) andCROW's (2016), active transport infrastructure or routes should be designed to be direct, cohesive, safe, attractive and comfortable.The poor condition of active transport infrastructures or routes, coupled with encroachments, inadequacy and their total absence in some developing countries have led to pedestrians and cyclists using major roads or lanes, unpaved pathways or footpaths (I-ce, 2000;Loo & Siiba, 2019;Tiwari et al., 2008;UNEP, 2016).In Accra, Tamale and Nairobi, cyclists were observed to have abandoned cycle lanes for motorised lanes and unpaved paths because the cycling facilities were either poorly designed or built (Green Africa Foundation, & UN-Habitat, 2014;Tefe & de Langen, 2008;UN-Habitat, 2010) whilst in some cases, they were either absent or inadequate (Acheampong & Siiba, 2018;Behrens & Newlands, 2022;Damsere Derry & Bawa, 2017;Nikitas et al., 2021).The majority of active transport users in developing countries, therefore, uses streets in lieu of inadequate and inefficient dedicated facilities (Nikitas et al., 2021;Tiwari et al., 2016).According to Fuseini and Kemp (2016), the streets that augment inadequate active transport infrastructure are themselves inadequate and unsafe due to rapidly increasing motorisation levels (Asamoah et al., 2013) and relatively high speed.The challenges associated with active transport infrastructure and the bottlenecks of the existing street for motorised vehicles are noted among the many reasons for the increasing skewness towards motorisation and the purported decline in walking and cycling (Timpabi et al., 2021).Walking is popular and dominant in Africa, and cycling is an important feeder mode for urban transport in most rural areas in the continent (I-ce, 2000).It can only be assumed based on the hugely noted active transport patronage, limited streets and footpaths and the challenged dedicated facilities like sidewalks and cycle lanes, that there are other undocumented routes or infrastructure supporting trips made on foot or by bicycles in the continent.Streets and cycle lanes or sidewalks, which are formal and known facilities, together with these unknown routes and infrastructures are giving people a reason to choose cycling or walking.These available options of routes or infrastructure for walking or cycling in Ghana are unknown, together with their extent of patronage.

Active transport infrastructure and demographics
Individual factors like gender and age have also been known to influence the choice of infrastructure.Females are said to evaluate active transport infrastructure differently than males and this influences their choice of infrastructure (I-ce, 2000).Female cyclists were found to prefer routes with maximum separation from motorized traffic to motorised roads with or without dedicated bicycle lanes (Carroll et al., 2020;Garrard et al., 2008).An observational study in Queensland showed that it was more likely for a cyclist to be a woman if they happened to be on roads with bike lanes, multiple traffic lanes or raised medians than on roads without these (Debnath et al., 2021).In India, pedestrian volumes on sidewalks were dissatisfactory to pedestrians, making females walk less frequently than males as they perceived a higher probability of collision or nearcollision incidents against male pedestrians (Saxena & Yadav, 2022).In an attempt to determine pedestrian preferences, Bernhoft and Carstensen (2008) found that older respondents appreciate pedestrian crossings and cycle paths significantly more than younger respondents.The demographic factors related to facility type and choices are useful for inclusiveness and avoidance of the incidence of abandoned facilities as observed in Ghana (Tefe & de Langen, 2008).It is important to identify present existing facilities and the role they play to active transport users, especially where infrastructure has been found to play an important role in active transport growth in other continents.It is of equal importance to investigate views regarding the use of these facilities to make informed policy strategies on infrastructure for promoting walking and cycling.This study therefore seeks to identify the various types of infrastructure or routes available for use by pedestrians and cyclists, as well as the extent of patronage juxtapose to formal infrastructures like sidewalks, cycle lanes and streets known to be already in use.The study will also determine the demographic relationship with the usage of these infrastructures and further solicit for views on the motive for the use of such facilities.Findings from this study will inform future policy decisions and planning of infrastructure for the growth of active transport in Ghana.

Study areas
As shown in the study framework in Figure 1, two study locations, namely: Tamale Metropolitan Assembly (TaMA) and the Ayawaso East Municipal Assembly (AEMA) were chosen for the study to represent the northern and southern parts of Ghana.The Ayawaso East Municipal Assembly may be the smallest municipality in Ghana with a total land area of 3.7 sq.km (AEMA, 2020).The AEMA is urbanized and densely populated with an estimated population of 118,688 people as of 2018 which represents 5.1% of the entire population of the Greater Accra Region (AEMA, 2020).The ratio of females to males is 1.04:1 indicating female dominance.The municipality has contributed to the fast-growing population of the Greater Accra Region and has a very youthful population with 54.72% of the population under the age of 24 years.The municipality houses two electoral and popular areas in Ghana, one of which is include Nima, a densely populated part with poorly planned settlement and transport infrastructure (Alaazi et al., 2021;Markwei & Appiah, 2016).Nima presents numerous challenges to socioeconomic infrastructure related to slum settlements, which represent a low-income status (AEMA, 2020).Some key transportation-related development challenges the municipality is identified with include street congestion from lack of access roads, bus terminals or stops, and inadequate street lighting systems.
On the other hand, the TaMA has a comparatively huge land size of 646.90180 sq.km making it the third largest city in Ghana (Tamale Metro Assembly, 2021).It is at the centre of the Northern Region of Ghana and has the central business district for trading and commerce with neighbouring countries like Togo and Burkina Faso (Ghana Statistical Service, 2014a).Tamale Metropolitan Assembly has a population of 950,124 per the 2021 Population and Housing (Tamale Metro Assembly, 2021) and about 80.8% of its population lives in the urban areas (Ghana Statistical Service, 2014a).The Tamale Metropolitan Assembly also has a youthful population with almost 36.4% of the populace below 15 years (Ghana Statistical Service, 2014a).Female and male percentages were 50. 3 and49.7, respectively, as of 2010 (Ghana Statistical Service, 2014a).There are networks of sidewalks and dedicated cycle lanes length of 51.92 km (Acheampong & Siiba, 2018) which are poorly developed (Damsere Derry & Bawa, 2017) and no match for the 251.6 km of the road network (ASIRT, 2014).There is therefore an obvious inadequacy in active transport users' facilities, which suggests that active transport users may rely on roads, of which about 47% are in poor condition (ASIRT, 2014).
The two locations, though very different in land size and location, have a youthful population, relatively high female dominance and similar ethnicity and culture.It is expected that AEMA will have a cycling culture due to the presence of migrants from TaMA (AEMA, 2020) as compared to other cities in the south of Ghana (Grieco et al., 1994).TaMA is already noted to have cycling as transport and its potential is said to be high in Ghana (ASIRT, 2014;Damsere Derry & Bawa, 2017;Timpabi et al., 2021).The study includes both cycling and walking therefore it is necessary to choose a place with not just walking which already is popular but also observed or purported bicycle

Literature review
Selection of two study areas: 1 North and 1 South, bicycle patronage

Questionnaire design:
Socio-demographics, trip patterns, perceptions, modes, facility types etc. URBAN, PLANNING AND TRANSPORT RESEARCH patronage, especially in the southern part of Ghana where the modal share for bicycles is said to be low (Acheampong, 2017;Quarshie, 2004).

Data collection design
The study was perception-based and used a cross-sectional semi-structured questionnaire survey at both locations to solicit views from pedestrians, cyclists and other road users.It also included screen-line traffic counts at selected locations in both study areas to establish the extent of usage of the formal facilities like sidewalks, cycle lanes and streets to verify responses of active transport users.Information on age, gender and occupation was sought from participants who also identified themselves as either active transport users or non-users when they confirmed whether they walked a bit on their trips, cycled a bit on their trips or depended on motorised vehicles (i.e.motorcyclists and private car owners or users).Walking and cycling in this context were explained to exclude egress or access to motorised vehicles within the confines of a home or building space.The age groupings were done to reflect on the occupation of respondents.Respondents below 16 years were expected to be at the compulsory basic education level (Aheto-Tsegah, 2011).
The age cohorts who were more than 24 years were assumed to be in the working class based on anecdotal evidence.Respondents who were more than 50 years were combined since their various groupings according to the local population pyramid (AEMA, 2020; Ghana Statistical Service, 2014b) or national one (Saroha, 2018) were less than 2% of the population.Under occupation, people were expected to choose the following: None (people with no form of occupation), Apprentice (any artisan under training), Student, Public worker (employees in a government facility) and Private worker (anyone working but not in a government facility and comprised of self-employed individuals, free-lance workers, farmers, workers in private enterprises, among others).Those who walked or cycled a bit were expected to choose a facility they frequently use from the options provided or name one under the option, 'other'.The facilities options included formal ones like the sidewalk, cycle lanes and streets and informal ones like unpaved footpaths and self-made routes within and around settlements, which are locally called 'lungulungus'.Participants of the survey provided their reasons or motivation for choosing a facility frequently.The GPS coordinates of respondents were recorded and displayed in QGIS 3.28.1 as shown in Figures 2 and 3.

Data collection
The questionnaire survey was undertaken using an open-source Android app for collecting survey data (KoBoToolbox, 2022a) called 'KoboCollect', which works through the KoBoToolbox, a suite of tools for field data collection (KoBoToolbox, 2022b).For each of the study locations, 700 questionnaires were administered, which was more than the estimated sample space of 600 estimated by the MaCorr (2020) equation at a confidence level of 95%, a z-value of 1.96, a 50% percentage picking a choice and a confidence interval of 0.04.Participants were sourced from the streets and in households by a face-to -face approach through simple random sampling.In TaMA, about 85% of the respondents were sourced at homes or activity points, and the rest were on the streets.In the Ayawaso East Municipality, about 81% of participants were sourced on the streets, and  the rest at destinations or homes.All participants were supposed to be people living in the study areas.Efforts were made to cover the entire AEMA whilst, in the case of TaMA, participants were interrogated mainly within and close to the Central Business District (CBD) where most of the populace is concentrated (Ghana Statistical Service, 2014a).To ensure full coverage, the survey was extended to some towns on the outskirts of the Metropolis as shown in Figure 3.
In addition to the questionnaire survey, 12-hour screen-line counts were undertaken on three routes at each of the study locations.More locations would have reflected more on the active transport share and number of other infrastructures but for available resources.The counts spanned from 6 am to 6 pm to capture the peak periods.The routes included double-lane dual-carriageway arterial roads, collectors and access roads.The locations of the screen lines had active transport infrastructures like sidewalks or cycle lanes beside the motorised streets to enable comparison of the active transport users' volumes on them.More screen line count locations would have given a better reflection of the active transport share and volumes on other infrastructure, especially in TaMA, which was relatively larger if not for limited available resources.Pedestrian and cyclist volumes on the available sidewalks or cycle lanes and the road carriageways were recorded every 15 minutes by enumerators who were positioned in both traffic directions of the roads.The locations of the screen lines for both study sites are shown in Figures 2  and 3 and they were chosen to cover places with high, moderate and low pedestrian or cyclist volumes.

Data analysis
Spreadsheet data was exported from the Kobo Toolbox and cleaned of inconsistencies.Six hundred and eighty-one (681) and 688 questionnaires were found to be valid for AEMA and TaMA, respectively.The valid data was analysed descriptively using the Statistical Package for Social Sciences (SPSS), a statistical platform (IBM, 2021).Inferential test in the form of chi-square test was also performed.The total 12-hour pedestrian and cyclist volumes on the main carriageway and those on the sidewalks or cycle lanes were determined and descriptively assessed together with the questionnaire data.

Active transport users and their characteristics
Out of the 681 and 688 valid respondents, 638 and 598 people mentioned that they engaged in a bit of cycling or walking on their trips in AEMA and TaMA, respectively, as displayed in Table 1.The rest of the respondents were mostly motorcyclists, private car owners or users and other motorised vehicle users whose active transport usage included only egress and access by walking or cycling to their motorised vehicle from home or destination point.AEMA recorded the highest percentage (94%) of active transport users, while the TaMA recorded 87% of the total respondents as active transport users.Interestingly, almost all the active transport users in AEMA were pedestrians.Four percent out of the 90% active transport users in AEMA used bicycles as a mode or one of their modes for their trips.Walking is therefore more dominant of the two active transport modes in both locations, especially in AEMA.The proliferation of motorcycles in TaMA reported by Fuseini and Kemp (2016) and Damsere Derry and Bawa (2017) may have contributed to the relatively high number of non-active transport users, leading to relatively low numbers of active transport users, especially pedestrians.Bicycle users in TaMA were about 15% of the valid respondents.TaMA is deemed one of the cycling hubs of Ghana (Acheampong & Siiba, 2018;Damsere Derry & Bawa, 2017) and has high bicycle ownership (Timpabi et al., 2021), hence a comparatively higher bicycle usage was expected as compared to AEMA.Generally, the number of respondents who indicated cycling at both locations was very low compared to the number who walked.This is consistent with the assertions by Acheampong (2017), and Loo and Siiba (2019) that cycling levels are low in sub-Saharan Africa.The active transport usage was assessed with some common and easily determined social demographics like age, gender and occupation, and per gender, males dominated the percentage of pedestrians and cyclists at both locations as shown in Table 1.The percentage of male pedestrians and cyclists in TaMA and AEMA were 38% and 10%, and 51% and 3%, respectively.Pedestrian respondents comprised mostly people who were between 25 and 35 years and those younger than 25 years.In TaMA, 33% of the valid respondents were pedestrians who were 25-35 years old and 12% were cyclists who were less than 25 years old.In AEMA 35% of valid respondents were pedestrians who were mostly between 25 and 35 years and a majority of the cyclists (2%) were at a much younger age than those in TaMA (they were mostly between 16 and 25 years).These age groups of pedestrians and cyclists in TaMA were mostly students and private workers.Students had the highest number of cyclists whilst private workers including farmers, artisans, private business people and people who worked in private firms dominated pedestrians at both locations.More than half of the 15% of cyclists in TaMA were students.A quarter of the few cyclist respondents in AEMA were also students.These findings re-echo the significance of students to programmes that seek to promote cycling as emphasised by Timpabi et al. (2021).Cycling programmes focused on students like the Ghana Bamboo Bikes Initiative (Whiting, 2020) are laudable and worth monitoring to enhance cycling and physical activity among students and consequently reduce the risk of cardiovascular and chronic diseases indicators reported to have increased over the years in Ghana (Kodaman et al., 2016).The promotional programmes on walking and cycling will also improve student productivity as observed in Kenya through their World Bicycle Relief programmes in 2015 (UNEP, 2016).
Further, a chi-square test of association was performed based on combined data from both study areas to investigate whether there is significant association between the respondents' socio-demographics and whether they cycle or walk a bit on their journey or are non-active transport users.For the purpose of this study, a significant association was considered at α ≤0.05.The findings shown in Table 1 revealed that although weak, the choice of mode of the respondents had a significant association with their sociodemographic characteristics including age, gender and occupation.Due to differences in culture and environmental conditions, a significant association between choice of mode and the study areas was expected.This was confirmed by the chi-sqaure test results: χ(2) =76.392, p=0.000, which showed that mode preferences of the respondents were significantly associated with where they live.Active transport infrastructure designers and transport planners may have to, therefore, take a critical look at demographics where there is the will to increase patronage of active transport.

Infrastructure type and motivation for usage
Respondents identified five infrastructure or route types frequently used.In AEMA, 81% of the 638 active transport users provided information on infrastructure or route types used frequently whilst in the case of TaMA 58% of the 598 active transport provided such information.AEMA has no cycle lanes and therefore no active transport user is identified with it as a frequently used infrastructure or route type.Out of the five route types outlined in Figure 4, the most frequently used in TaMA was "the self-made routes within and around settlements commonly called 'lungulungus'.Thirty-six percent (30% pedestrians and 6% cyclists) of the 345 respondents said they used 'lungulungus' in TaMA, followed by those who frequently used streets, who formed 25% of the 345 active transport respondents.Twenty-three percent (23%) of the active transport users in TaMA used sidewalks, which is almost the same number as those who used streets, making sidewalks as popular as streets in TaMA.Notably, in Ghana sidewalks or walkways are usually placed along the streets (Amoako et al., 2014) and therefore it can be expected that usage of both will be similar or the sidewalks will be more patronised by active transport users.In TaMA, the cycle lanes also serve as sidewalks.In terms of cyclists' usage of cycle lanes, only 3% of the 345 respondents, or 11% of the 101 cyclists, said they used cycle lanes as their infrastructure, whereas the majority of respondents who comprised 43% (30% pedestrians and 13% cyclists) relied on the street as their infrastructure in TaMA.This finding buttresses the observations made by Tefe and de Langen (2008), UNEP (2016) and Loo and Siiba (2019) on cyclists using streets other than the provided cycling infrastructure.
In AEMA, motorised vehicle streets were the most frequently used infrastructure followed by the 'lungulungus', then footpath (unpaved).Thirty-seven percent (37%) of both cyclists and pedestrians frequently used streets (motorised) in AEMA followed by 32% of the active transport users who used 'lungulungus'.The least used infrastructure in AEMA was sidewalks, which 13% of the active transport users identified as their most frequently used infrastructure.Most pedestrians used streets (34%) more frequently than they did sidewalks (12%), or unpaved footpaths (18%), even though the patronage of sidewalks by pedestrians in AEMA was more than those in TaMA.Ninety-one percent of those who used sidewalks in AEMA were pedestrians whilst TaMA had 76% of sidewalk users being pedestrians.It was surprising that sidewalks were not as well patronised as streets were in AEMA and TaMA despite their being notably aligned along the streets (Damsere Derry et al., 2010).Generally, there was relatively low patronage of sidewalks and cycle lanes at both locations as compared to the other infrastructure or route types.The low patronage of sidewalks and cycle lanes may be due to the lack of them as found in the case of cycle lanes in AEMA (AEMA, 2020;Owusu et al., 2008) or their inadequacy as noted by Naami (2019).Buehler et al. (2016) and Amoako et al. (2014) mentioned that the aforementioned reasons were part of the obstacles to walking and cycling.The low patronage of cycle lanes and sidewalks may also be a result of design flaws (Tefe & de Langen, 2008;UNEP, 2016) or just out of sheer disinterest on the part of users which is commonly noted of infrastructure provision that fail to incorporate stakeholder's and users' views (I-ce, 2000).
In terms of demographics of respondents, chi-square test of association was also performed based on combined data from both assemblies to investigate whether there is a significant association between the socio-demographics and the choice of facility usually used.A significant association was considered at α ≤0.05.The test of association between gender and choice of route was found to be statistically not significant at χ(4) = 5.136, p = 0.274.This means that infrastructure preference had no association with gender.The preference of males do not, therefore, differ significantly from that of females when it comes to infrastructure or route type choice for walking or cycling.Gender may, therefore, not influence infrastructure or route choices.This was observed in the number of respondents where majority of both male and female active transport users in TaMA indicated that they used 'lungulungus' as their infrastructure frequently.However, more females in TaMA mentioned using 'lungulungus' (i.e.self-made routes within settlements) than males.Fifteen percent (15%) and 21% of the active transport respondents who are males and females in TaMA, respectively, said they used 'lungulungus', frequently.'Lungulungu' routes are convenient routes within settlements defined by people.Since they are not aligned along the streets, they may be deemed safer from vehicular interactions.Safety is a characteristic of active transport infrastructure or routes known to attract females (Carroll et al., 2020;Garrard et al., 2008) hence the relatively high female usage of 'lungulungus'.In AEMA, the similarity in infrastructure usage by both males and females was observed.The proportion of female active transport users who used streets and 'lungulungus' were the same and formed the majority, as shown in Figures 5 and 6 (for TaMA) and most males who represented 21% of the 345 active transport respondents used streets.
Unlike gender, a statistically significant association was found between the age and route choice of active transport users (χ(16) = 49.152,p = 0.000).The majority of the older age groups of active transport users who were above 16 years in TaMA used 'lungulungus' frequently whilst those below 16 years surprisingly preferred to use streets frequently as their infrastructure despite the high vehicular speeds and unsafe conditions.The choice of route was also found to be significantly associated with the type of occupation (χ(16) = 57.502,p = 0.000).More than half of those who used streets as their main infrastructure in TaMA were students whilst most of the working group including both private and public workers used mostly 'lungulungus'.In the case of AEMA, 13% of the 516 active transport respondents who were students used lungulungus.Among the working group, the most used streets were 21% of active transport  respondents.The students and working groups belonged to mostly the 16-24 and the 25-35 years groups.Notably, those more than 50 years in TaMA used the unpaved footpath more than the other routes.Out of the 3% active transport respondents who were more than 50 years, 2% indicated using footpaths frequently.The footpaths and 'lungulungus' are usually separated from motorised vehicles, making them safer for the older group who consider safety as a determinant of walking or cycling (Bernhoft & Carstensen, 2008).Footpaths, however, were the only available facility for those who indicated that they used it frequently in TaMA, which underpins the inadequacy of dedicated infrastructure making cohesion (availability) a deficient character of dedicated infrastructure in TaMA.
As expected, a significant association was found between the active transport users from the two municipal assemblies with regard to their choices of facility from the chisquare test: χ(4) = 50.916,p = 0.000.This could be due to differences in conditions of the infrastructure or route types, environmental conditions, as well as travel culture at the two locations which may introduce variations in the preferences of the active transport users from the two cities.
At both study locations, the majority of the active transport respondents by age, gender or occupation thought the infrastructure they chose to use gave them the shortest route option.The motivation for shortest routes re-emphasises the desire of active transport users, especially in the metropolis, for shorter travel time, as noted by Timpabi et al. (2021).They comprised 67% and 63% of the 345 and 516 active transport respondents at TaMA and AEMA, respectively, from Table 2.The directness of routes was found to be the premium consideration of active transport users in their choice of infrastructure.Directness looks at reducing delay by providing shorter routes with no or few detours (CROW, 2016).Walking and cycling use their users' energy (CROW, 2016;  URBAN, PLANNING AND TRANSPORT RESEARCH Tiwari et al., 2016).It is therefore more desirable if routes or infrastructure provide shorter paths for users.Directness should therefore be the primary design criteria for active transport infrastructure planning at these two locations and in Ghana at large to ensure patronage of infrastructure and active transport.
In TaMA, most pedestrians and cyclist respondents, aside from looking at shorter routes in their choice of infrastructure, secondly mentioned that their infrastructure choice was based on the fact that they had no risk of accidents.Sixteen percent (16%) of the 345 active transport respondents in TaMA said their frequently used infrastructure or route choice has no risk of accidents.It can be implied that safety is the second primary concern of people when it comes to active transport infrastructure in TaMA.The absence of risk of accident as a motivation underscores the high usage of 'lunlulungus' in TaMA.'lunlulungus' are usually within settlement areas and separated mostly from motorised vehicle streets.In essence, totally segregated or separated dedicated infrastructure will be preferred in TaMA.In AEMA, besides shorter routes, active transport respondents chose to use streets more than any infrastructure because they deemed them the only available facility.Twenty percent (22%) of the 516 active transport respondents mentioned this, making it the second important reason for infrastructure choice and subsequently putting cohesion, which looks at availability and continuity, as an important infrastructure feature in AEMA.In the absence of or limited dedicated infrastructure, as seen in AEMA, people are left to use the streets.In terms of the dedicated infrastructures like the cycle lane and sidewalks, they were chosen by their users because they thought it provided them first with the shortest route option, have no risk of accident or were the only available facility.The perception of no risk of accidents associated with sidewalks or cycle lanes in TaMA may be a result of the separation of these infrastructures from the streets they are along by kerb delineators as observed.
The same trend of motivation for the choice of infrastructure is observed across the various gender, age and occupation groups as shown in Figures 7 and 8.In TaMA, however, a high number of females in addition to considering the shortest route and no risk of accident in infrastructure choice also mentioned that they considered 'lungulungus' and other route types because they were either the only available ones (3%) or the only dedicated ones available to walk or cycle on (3%).This indicates that the availability of the facility or cohesion is also important to females.It is important to identify the views of the inhabitants for location-based planning of infrastructure and not a one-fits-all method, which leads to the abandonment of infrastructure.
There is a huge active transport user reliance on streets and 'lungulungus' other than their designated infrastructures such as sidewalks and cycle lanes.A 12-hour screen-line count of active transport users on collector, access streets and a major arterial road at each of the study locations confirmed the findings that streets are more patronized by active transport users, especially cyclists, as infrastructure than the sidewalks or cycle lanes.The volumes of cyclists were generally lower, which confirms the observation made by Loo and Siiba (2019) of African countries.At the three screen-line locations in AEMA, the percentage of cyclists riding on streets was more than those using the adjourning sidewalks.Location A1, A2 and A3 had 81%, 62% and 85% of their cyclist volumes using the street carriageway, including that of major arterial or highways, as their infrastructure, as shown in Table 3.This reiterates observations made by Tefe and de Langen (2008) and Loo and Siiba (2019) that cyclists use streets even when there are provided facilities because the facilities are inadequate, flawed in design or poor in appearance.Similarly in TaMA, two of the locations: T2 and T3 had 76% and 98% of their total  URBAN, PLANNING AND TRANSPORT RESEARCH cyclists' volumes using the street carriageway.At T1, more cyclists used the cycle lane, which doubles as sidewalks rather than the carriageway, probably because the latter have kerb delineators separating them from the highway they are aligned along and provide a sense of safety, or it may be due to high-speed vehicles, which pose a risk to users in the carriageway.At both locations, pedestrians resorted to using sidewalks or cycle lanes more than they did the motorised roads.AEMA had the highest percentages of pedestrians using sidewalks on the various routes.Pedestrian volume on sidewalks at AEMA was 94% and 93% at A1 and A2 locations, respectively, and 99% and 64% at T1 and T2, respectively, in TaMA, which confirmed the relatively high pedestrian respondents who   mentioned that they used sidewalks in AEMA from the survey.Places where most pedestrians used streets were A2 and T3 where sidewalks were observed to have poor surfaces and conditions and encroachment by trading activities.The findings from the count are consistent with the survey outcomes, especially where cyclists are concerned.The dependence of pedestrians and cyclists on motorised streets suggests that the active transport facilities may either be inadequate (Acheampong & Siiba, 2018;Acheampong, 2017;Damsere Derry & Bawa, 2017;Lall et al., 2017;UN-Habitat, 2010), non-cohesive, poorly designed or absent (Tefe & de Langen, 2008) or in poor condition (Naami, 2019).These bottlenecks in active transport infrastructure are a common observation in sub-Saharan Africa and developing countries (Nikitas et al., 2021;Tiwari et al., 2008).It is important to know the preferences or choices of active transport users and views to be able to support the already dwindling and declining use of the foot and bicycles for transportation.The incorporation of the views of users addresses their needs thereby maximizing the use of dedicated facilities and improving existing ones through proper infrastructure planning and policy statements, geared towards active transport growth.Walking and cycling have been given attention as sustainable modes globally.In the Netherlands, where major promotion and growth successes are known and observed in walking and cycling, the density and continuity of active mode infrastructure are very high (Ton et al., 2019) making distances shorter and improving cohesion.From the study, characteristics, such as directness (shorter distance options), low risk of accidents (safety) and cohesion (availability and continuity), were again identified as the driving force in infrastructure choice and should be used as design criteria for infrastructure planning.This may reduce the use of informal routes like 'lungulungus' or walking or cycling through peoples' properties or within settlements and using streets, which results in high pedestrian and cyclist casualties.

Conclusion
A study was undertaken to investigate infrastructure and available routes for respondents who choose to walk or cycle on their trips in the Tamale Metropolitan Assembly and Ayawaso East Municipal Assembly in Ghana.First, socio-demographics including age, gender, and occupation all had significant association with their mode choice.Second, respondents at both locations identified five route choices or infrastructure, which included mixed dedicated infrastructure for pedestrians and cyclists (sidewalks and cycle lanes) in the Tamale Metropolitan Assembly and sidewalks in AEMA.In addition to the aforementioned infrastructure, respondents mentioned using streets, unpaved footpaths and self-defined routes around and within settlements, which are popularly and locally referred to as 'lungulungus'.Among these infrastructures or route choices, the majority of the respondents used 'lungulungu' routes and streets in TaMA and AEMA, respectively.These two, according to the respondents, provided them with the shortest routes, making directness an important infrastructure feature for planners and local governments to consider in the design of active transport infrastructure in Ghana.Sidewalks were the most frequently used dedicated infrastructure by both pedestrians, who formed the majority of the active transport users, and cyclists, who were very low in numbers.Apart from gender, socio-demographics of active transport users including age, occupation, and area of abode were all significantly associated with the choice of facility type.Across the various age groups, gender and occupation, streets and 'lungulungus' remained the frequently used routes.The high use of 'lungulungus' indicates that the built environment is poorly planned.It also suggests that the streets that are noted to be also highly used and the dedicated infrastructure for walking and cycling are not available, adequate or direct which would have resulted in people preferring to manoeuvre around and among settlements to get to important places like schools and work.Directness was the topmost criterion for the choice of an infrastructure to walk or cycle on at both locations.Safety and cohesion (availability) were the next important criteria for infrastructure choice in TaMA and AEMA, respectively.If higher volumes of cyclists and some pedestrians preferred to risk their lives on high-speed highways or major arterial roads at both study sites rather than using the dedicated facilities provided right along these roads, then it suggests that the facilities do not meet their users' preferences, which underpins the low patronage of the facilities and low active transport patronage.
The study is useful to inform plans and policy decisions on active transport infrastructure as it incorporated respondents' views and needs to reduce the incidence of infrastructure abandonment, which is observed in Sub-Saharan Africa.Existing active transport infrastructure should be technically reviewed by local planners and designers with directness, cohesion and safety as some major criteria since these were the drivers for route and infrastructure choices for walking and cycling.Future active transport infrastructure should be planned with these criteria and local preferences that will drive the choice to walk or cycle towards promotion and growth in active transport in Ghana.Despite these challenges bedevilling active transport modes and infrastructure setbacks, especially with cycling in Ghana, students' patronage was high.They are, therefore, highly recommended for programs geared towards promoting active transport, especially cycling, organised by the central government, local and school authorities and cyclist groups.
It is worth mentioning that this study adopted mainly descriptive statistics and inferential statistics in the form of chi-square test of association to investigate the patterns and relationships between sociodemographics and active transport trip characteristics.The findings of the chi-square test of association are not conclusive to make causal inferences concerning the data because the groups being analysed are not homogenous -there is no control for other variables.It is therefore recommended in future studies to employ analysis methods such as regression modelling that explains the complex interactions among variables.

Figure 2 .
Figure 2. A map showing respondents' locations and traffic census points within the AEMA study area (source: created by the authors from QGIS 3.28.1).

Figure 3 .
Figure 3.A map showing respondents' locations and traffic census points within the TaMA study area (source: created by the authors from QGIS 3.28.1).

Figure 4 .
Figure 4. Infrastructure or route choices by active transport users (peds and cyclists) in (a) AEMA and (b) TaMA.

Figure 5 .
Figure 5. Infrastructure or route types and demography of users in AEMA.

Figure 6 .
Figure 6.Infrastructure or route types and demography of users in TaMA.

Figure 7 .
Figure 7. Motivation and demographics of respondents' choice of infrastructure or routes in AEMA.

Figure 8 .
Figure 8. Motivation and demographics of respondents' choice of infrastructure or routes in TaMA.

Table 1 .
Active transport users' and non-users' characteristics.Values for Tamale Metropolitan Assembly are in brackets () and those for Ayawaso East Municipal Assembly are without brackets.2. Chi-square test of association results are based on data from both study areas combined.

Table 2 .
Motivation for infrastructure choice by cyclists and pedestrians in Ayawaso East Municipal Assembly and Tamle Metropolitan Assembly.Note: Values for Tamale Metropolitan Assembly are in brackets () and those for Ayawaso East Municipal Assembly are without brackets.

Table 3 .
Twelve-hour volume count of pedestrians and cyclists on either streets or dedicated infrastructures in AEMA and TaMA.