Bird composition of different valley habitats after land-use changes in Northern Honduras

The northern coast of Honduras is potentially covered with tropical rainforests, reaching from the Caribbean Sea up to the cloud forests of the Pico Bonito summits. Therefore, it was blessed with the megadiverse avifauna of the Central American humid neotropics. Although local bird species have been generally well documented, there are hardly any updates on the biodiversity of northern Honduras. Thus, this study contributes to our knowledge of the natural shift of bird life, following up the Cangrejal River with its different slight land use intensification in the region. Standardized bird records along the valley are analyzed, reaching from the beaches of La Ceiba up to the managed rainforests of El Toncontíns in the lower montane rainforests. Nine points were checked over the course of at least 6 days, taking point counts between 16 March and 20 June 2005. A NMDS of the joined nine point-lists elucidates four main groups, namely the beach/city ecosystems, open habitats along the river banks, slightly cleared forests (park landscape), and a mature rainforest. In total, 115 bird species, from 102 genera and 44 families, were found in 2005. As methods are limited, results can represent merely a prodromus of bird composition of neotropical valleys of the Central American isthmus. However, avi-diversity is affected by forest degradation and increasing land-use changes. Since deforestation is still soaring in the region, bird species composition should be monitored, as it will be as dynamic as land use changes in the region.


Introduction
The northern Caribbean coast of Honduras belongs to the per-humid tropics (Lauer et al. 1996), is therefore potentially covered with evergreen rainforests (Holdridge along the river habitats and to which altitude the protected bird species from the national park would descend and enter into the more disturbed areas at the bottom of the valley. Thus, although the Río Cangrejal is frequently visited by naturalists and birdwatchers, monitoring of the region and publication remains poor (Marcus 1983;Frederick et al. 1997;Anderson et al. 2004). Even more, tourists and scientists have been threatened by a growing number of crimes in Honduras in recent years, so that information on biodiversity might get scarcer still. Thus, the purpose of this paper is to report about some bird records, to provide a systematic prodrome of an avifauna of a steep Central American valley, and to document species shifts due to modest land-use changes.

Methods
Audio-visual bird observations were carried out along the Río Cangrejal valley in the hinterland of La Ceiba, department Atlántida, northern Honduras. Bird records of 28 observation days, between 16 March and 20 June 2005, were carried out, using a 10 × 42 binocular. Birds were mainly checked along the road "Carretera de la Cuenca" that leads from La Ceiba south along the river up to Río Viejo. One examined branch of the road follows the Río Yaruca south-eastwards to the villages of Yaruca, El Toncontíns, and El Urraco ( Fig. 1; Table 1). Bird lists were set up at nine points (Pa-Pi) which were located approximately 2 to 3 km apart from each other. The beach (point a= Pa), i.e., from the jetty "Muelle Turistico Reynaldo Canales" to the western estuary of the Río Cangrejal river, and the city centre of La Ceiba (172,000 inhabitants in 2007; Brinkhoff 2020) have been visited on eight   Fig. 2. A few further noteworthy single observations are stated more in detail, giving the date, location and a special remark. Further status information on each species is added according to Beall (1997), Frederick et al. (1997), Bonta and Anderson (2002), Ordoñez and House (2008), and IUCN (2020). The order of bird classes is used as in Chesser et al. (2017) and Gill et al. (2020), the order of families and species according to the Honduran species lists (Dickinson and Remsen 2013;Dickinson and Christidis 2014;Collar 2014, 2016;Lepage and Warnier 2020). For the analysis of species composition of the different points, presence and absence data records were used to conduct a NMDS-analysis. It was done by using the "vegan" package of the statistical platform "R" (R Core Team 2013; Oksanen et al. 2019). According to the NMDS, the nine species list of the respective study points are therefore assigned into four main groups (p < 0.0001).
For the two land-use maps, we used Landsat 7 ETM+ images of April-May 2005 and May 2018. Due to an error of the Landsat 7 sensor since 2003, there is a failure of the scan-line-corrector (SLC), leading to line-shaped data gaps (Maxwell et al. 2007). These black stripes are found in 20% of all Landsat 7 satellite images, and were filled up with images of subsequent years (2006, 2007, and 2019). Categories were just roughly split into four classes to detect deforestation in the region (water, forests, pasture/fields, settlements).

Results
During the study time, 115 bird species, belonging to 102 genera and 44 families and 15 classes were reconfirmed in the Río Cangrejal watershed in this study. Table 2 illustrates the total bird list, providing also some information on the conservation status and all localities of the respective species.
Thirty-seven species were recorded at La Ceiba (Pa), whereas 41 (Pb), 36 (Pc), 43(Pd), 45 (Pe), 38 (Pf), 44 (Pg), 31 (Ph), and 33 (Pi) species were found at the respective other points, resulting in a mean diversity of 38.7 species/ point (SD 5.1).   Principe (1999). A= accidental; N= native, breeding; V= visitor, non-breeding; W= widespread visitor, locally breeding; 1= common; 2= uncommon. NT= nearly threatened according to Bonta and Anderson (2002) and IUCN (2020). *= out of its common range. Point counts: Study points from Pa until Pi (a-i), where the respective species was found. One individual has been discovered in an unusual habitat of the city centre of La Ceiba at the Av. Morazan/ Calle 15, on 20 June at noon. Fig. 3 shows the NMDS of the nine species lists, illustrating a significant differentiation into four main groups-the city/beach group (G1), the river banks (G2), the degraded forests (G3), and the rainforest group (G4; Fig. 3). It also elucidates the isolation of G1, which can be explained by the manifold waterfowl and migrating birds such as H. rustica and Pandion haliaetus (Linnaeus, 1758) which are migrating westwards along the coast, to North America during spring time. These bird species do not enter often into the valley and do not exceed Pb (Fig. 3). However, some waders like the two Yellowlegs enter up into the lower Río Cangrejal valley, but Pa still remains quite different. The main group formed by the NMDS is G2, comprising also Pd. The group can be explained by the similar disturbed vegetation types along the river banks. Thus, secondary vegetation, gardens, shrubby pastures, and other open habitats with scattered trees or grassy river banks bring about the typical bird composition of open-habitats. As groves will be closer in the middle of the valley and show rainforest-like vegetation structures, biodiversity enhances, also resembling Pi. Pf is located 200 m higher than other points of G2, as the road is leading over a pass and is therefore providing much more forested habitats. Pg and Ph are different from Pa-Pe as the river gets another character, by carrying much more, and wilder, water (white-water tourism) than above the river confluence near Río Viejo. This might also be the reason that Phalacrocorax brasilianus (Gmelin JF, 1789) is no longer found south of the constriction of the valley. For the same reason, Ph seems to be most different in G2, as the eastern river branch is carrying less water than the main Río Cangrejal. Pi assembles the bird observations of the upper valley around El Urraco and the timber huts in El Toncontíns and reflects the higher biodiversity of the more intact and nature-like rainforest habitats, referring to G3 that is also less disturbed but also incorporated pasturelands and tree islands.

Discussion
The results reflect the expected impoverished diversity in comparison to intact rainforests (Adams and Ruiz 2017; EBird 2020) and national parks, where up to 435 bird species are found at one location (Lepage and Warnier 2020). They show the uniform composition of open habitats and disturbed forests along river banks and the rather quick shift from the sea to the lower mountain rainforests of a steep Caribbean valley. As the region suffered from a moderate alteration into a tessellated landscape, it also includes the well-known invasion of common neotropical generalists (Hohnwald 2009;Moura et al. 2013). However, some bird species come from natural surrounding forests at the hill sites, particularly Streptoprocne zonaris (Shaw, 1796), Tachyphonus luctuosus (d'Orbigny & Lafresnaye, 1837), and Ramphocelus sanguinolentus (Lesson, 1831). But no indicator species of cloud forests like Pharomachrus mocinno (de la Llave, 1832), Geotrygon albifacies (Sclater PL, 1858), or Xiphorhynchus erythropygius (Sclater PL, 1860), that are supposed to live at the summits of the Pico Bonito mountain ridges (Howell and Webb 2001;Gallardo et al. 2015), were observed.
As the intention of this study was neither to detect new species for the region nor to complete species lists of special or unique habitats but rather to document the scarcely studied species of the remaining degraded landscapes, species numbers must remain low. Therefore, the slightly to heavily disturbed environments, studied rather from a geographical point of view, do not provide any rare species, or further surprises, or extraordinarily high count levels. Thus, the results stand in contrast to intensive birding that often includes mist-netting, microphones and other intensive methods Loiselle 2000, 2001;Lepage and Warnier 2020). However, point counts are known to be a valuable complement to mist-net studies, if certain methodological pitfalls are considered (Blake and Loiselle 2000;Robinson et al. 2018). Other reasons for the low species numbers might be the limited time frame with missing night trips or the fact that remaining habitat patches are too small and subdivided to sustain a more diverse avifauna (Young et al. 1998). Another technical reason might be that point-count-methods are less appropriated in the humid tropics in general, and they might need other standards than in Central Europe (Fischer et al. 2005). Bird counts here were recorded between field work of a forest project in the Río Cangrejal watershed (Kukkonen et al. 2008;Kukkonen and Hohnwald 2009) and often there was no time to track down calls or to follow shy bird species (Robinson et al. 2018). Thus, species accumulation curves in Fig. 2 elucidate that especially the bird diversity of older rainforests, for instance Pi, Pf, and Pe, are not yet representatively acquired. However, several curves (e.g., Pc-Pd) suggest soon approaching saturation levels. At least, 115 of the 604 detected species in the department Atlántida, and 363 of the National Park Nombre de Dios and 401 of the Pico Bonito national park (AVIBASE 2020) could be reconfirmed for the valley and two of "accidentally" occurring species (status A in Table 2) reaching its northern distribution frontiers. At least five species were registered to a small extent out of their known main distribution range (Howell and Webb 2001;IUCN 2020; Table 2). In all, it is concluded that the presented data are sufficiently reliable to perform a prodromus of the bird diversity and species composition of a rather steep neotropical valley of the Central American Isthmus.
Although the applied methods are surely just about comprehensive enough to detach the commonest bird species at the locations (Gotelli and Colwell 2001;Robinson et al. 2018), the NMDS unambiguously shows the different species compositions within the lower Río Cangrejal valley during spring time (March-June; Fig. 3). Species lists will certainly be easy to be stocked up by intensified methods and extending updates from the valley can be verified at the internet platform eBirds (EBird 2020). There, meanwhile, 15 points along the Río Cangrejal valley with mainly added up species lists are well documented. Points with up to 15 checklists are comparable with the findings of this study. However, there is amazing yearround monitoring of up to 130 observation days, counting 216 bird species in the middle of the valley. High counts are the Río Cangrejal delta with 239 bird species from 485 year-round checklists, 216 species at the Omega jungle lodge (130 expert lists), 151 at the Pico Bonito national park (Visitors Centre, La Roca trail; 32 added up checklists), 138 species at Las Mangas (with just 10 expert lists), 132 species of the Pico Bonito national park (section El Naranjo, 16 expert lists), and 132 species at the CREDIA Botanical Garden in La Ceiba (with year-round 337 checklists; EBird 2020). However, one advantage of this study is the rather systematic approach with a fixed design that can be easily replicated in subsequent years.
However, expanding the lists by experts or spending more time or putting microphones at the respective points will not necessarily decrease similarities between the central valley points (Pb-Ph). The contrary is more probable: if more observation time is spent at all points, the number of common species certainly will also increase. Just in the cases of the Pa and Pi, differences surely will enhance with intensified sampling methods (arrows in Fig. 3). Comparing these lists with the results of this study is problematic, as methods are different and time and knowledge of the observers play an important role (Robinson et al. 2018). But results resemble data from intensive elevational studies in Costa Rica, where habitat disturbance and size are as important as altitude (Young et al. 1998;Loiselle 2000, 2001). Comparing the study sides (Cangrejal, Tilarán, La Selva, Braulio Carillo) elucidate many common characteristics such as the lower mountainous environments, the Caribbean eastern slopes, and the humid tropical life zones. However, there are also some ecological differences in the settings of the inventories, as they were sustained by intensive mist-netting and were located at slightly higher regions with less anthropogenic disturbances (Tilarán 650-1,700 m a.s.l.; La Selva 50-500 m, and 500-2,000 m Braulio Carillo). Therefore, these studies could be seen as an elevational continuation of this study. However, mountains in Honduras reach much higher maximum altitudes and are located 630-680 km further NNW. Other differences are that the Costa Rican study regions are located far away from the Caribbean coast (77-135 km) and thus completely lack the sea waterfowl diversity.
There are some ecological gradients within the data which might have some effects on bird compositions. For instance, there is a north-south gradient, thus a humidity as well as a slight hypsometric gradient that might explain some minor species shifts. However, passing the Río Cangrejal watershed, semi-arid valleys with a much more xeric vegetation replace the rainforest biome and might be quite different (Holdridge 1962). Thus, patchy land use patterns of old forests, secondary forests, pastures, fields, town areas, and its different grades of habitat naturalness might have much higher impacts on species compositions than any other ecological gradient. There had been some short day-trips into the national park and the mountain rainforests during the forest project but for the lack of time, there was no chance to carry out bird surveys. The same is true for the Cangrejal delta, so that important data for the complete species shift of the lowest and uppermost valley are still missing which would be interesting for future studies. Forest field work with horses did also not allow for the detection, or follow-up, of shy, rare, and nocturnal bird species. Additionally, surveys of the summits would be also of great interest to complete the biodiversity patterns.
Different swallow species seem to avoid using the same habitats, as H. rustica, e.g., was just observed at the coast, while Stelgidopteryx serripennis (Audubon, 1838) stayed in the upper parts of the valley and Riparia riparia (Linnaeus, 1758) just at the lower valley (Pb). The four individuals of Petrochelidon pyrrhonota (Vieillot, 1817) appeared at El Naranjo (Pd). There was an example of reported mass migration of H. rustica in northern Honduras (Anderson et al. 2004). However, own observation here showed very even migration of averagely 10 single individuals/ minute, along the coast.
Management of FSC-certificated forests, but actually also conventional managed forests, is carried out in extensive and ecologically acceptable ways (Kukkonen et al. 2008;Kukkonen and Hohnwald 2009; Bieri and Nygren 2011), which is also convenient for birds and other local wildlife. Thus, during field work, also Boa imperator Daudin, 1803 (Boidae, ca. 1.8 m long) and Atropoides mexicanus Duméril, Bibron & Duméril, 1854 (Viperidae, photo available) were observed in these forests (cf. Marineros 2000). For nature conservation, it is important to maintain or install biocorridors within the Cangrejal valley, especially between the two national parks to guarantee gene flows among populations. For most bird species, the extensively used landscapes are no further problem to pass or overfly. Understory birds like, for instance, some forest tinamous (Tinamidae), quails (Phasianidae) and rails (Rallidae) but actually also for ground mammals (e.g., Panthera onca and Leopardus pardalis both Linnaeus, 1758), forest amphibians and reptiles etc., would profit from several small forest corridors. The lower valley, with its secondary vegetation, is less useful for the species, as the busy city is too close and the road until Las Mangas too continuously settled. Thus, the best option would be a biocorridor south of the Cangrejal bridge, between El Pital and Río Viejo (between 15°40'16"N, 86°42'06"W and 15°39'58"N, 86°41'21"W), as the road leads up into the mountains, getting narrower and apart from the river, so that animals could pass over more easily the two alleged obstacles.
However, outside the managed forests, deforestation activities expanded in slow but continuous rates, in the whole Cangrejal watershed, especially in the south-east of the Cangrejal region (Fig. 1). The figure also elucidates, along with the much higher resolution in 2018, that deforestation has entered both national parks, which seems to be a quietly accepted habit in the valley. For conservation issues this might be the most critical point and governments should clarify and control land titles, and strictly insist on once established national park borders. Further, the higher resolution also elucidates that within the agricultural areas, there are many smaller forest patches that might maintain a considerable part of the biodiversity in the landscape or can deal as stepping stones for animal migration. However, deforestation soars in the region and the future of indigenous forest biodiversity in northern Honduras has to be seen with concern. But as the rich bird diversity might attract more ecotourists and birdwatchers, the unique natural resource "biodiversity" might get more important in future. Local agricultural production hardly produces enough income to make a satisfying living in the region (Nygren and Myatt-Hirvonen 2009), and ecotourism might be one idea to improve the smallholders' situation or, at least, to valorize their agricultural products. However, tourist infrastructure and safety in northern Honduras would then have to be improved. For nature conservation, all the aforementioned issues should be discussed with politicians, NGOs, landowners and smallholders, to create a sustainable joint future for the people and bird life of the beautiful and unique Río Cangrejal valley.