Urban vegetation change after a hundred years in a tropical city (San Jos\'e de Costa Rica)

Urban vegetation is of key importance because a large proportion of the human population lives in cities. Nevertheless, urban vegetation is understudied outside central Europe and particularly, little is known about the flora of tropical Asian, African and Latin American cities. We present an estimate of how the vegetation has changed in the city of San Jos\'e, Costa Rica, after about one century, with the repeat photography technique (based on a collection of 19th and early 20th century photographs by Jos\'e Fidel Trist\'an and others) and with data from the Costa Rican National Herbarium. We found little vegetation change in the landscape of San Jos\'e during the 20th century, where a total of 95 families and 458 species were collected in the late 19th and early 20th century. The families with most species were Asteraceae, Fabaceae, Poaceae, Lamiaceae, Euphorbiaceae, Solanaceae, Cyperaceae, Acanthaceae, Malvaceae, Piperaceae and Verbenaceae. Similar results have been found in Europe, where the number of plant species often is stable for long periods even when the individual species vary.

Urban vegetation has mostly been studied in central Europe, where about 50% of species are alien, half of them introduced before the 15 th century. Despite the heavy traffic among European cities, five centuries have not been enough to homogenize their flora: the communities of species introduced after the year 1500 are characteristic of each city (Frank et al. 2008).
For animals, which particular species occur in cities is predicted by the "environmental filtering model" that in turn is based on plants. The model states that (1) there is natural selection of species living in urban ecosystems, (2) plants define key habitat characteristics and (3) habitats define which animals can live in the city (Williams et al. 2009). Generally, moderate urbanization produces some increase in plant biodiversity but is deleterious for invertebrates and mammals. A high level of urbanization is correlated with fewer species of plants, invertebrates, amphibians, reptiles, birds and mammals, possibly because humans willingly introduce plant species, but not animals, to their gardens (McKinney 2008).
Outside central Europe, urban vegetation is understudied but there are some recent data from Plymouth, England, where alien species increase with urbanization (Kent et al. 2001). Also in England, gardens in Sheffield have a total of 1 166 plant species (70% alien) and twice the garden size means 25% more species. In these gardens there are 63% biennials/perennials, 18% shrubs, 10% annuals and 8% trees (Smith et al. 2006).
In Anglosaxon and French North America, there is a surprising scarcity of recent studies on urban biotas, but some work has been done.
In the city of Halifax (Nova Scotia, Canada), soil moisture and light determine which species are present. Taxa adapted naturally to rock, grassland and flooded habitat find an analog habitat in the city and thrive (Lundholm & Marlin 2006).
The New York metropolitan region has 556 woody species and non-native invasive species are becoming more common (Clemants & Moore 2004). In the Pelham Bay Park, New York City, native species went from 72% to 60% and 26% of natives disappeared in 50 years (especially herbaceous and meadow-type plants, DeCandido 2004).
Even though recent studies are scarce, a meta-analysis found 79 studies of species richness with geographic data for New York City; of these, 17 studies found a decrease in species richness, six an increase and three found no change. However, all studies reported an increasing number of exotic species (Puth & Burns 2008).
Tropical cities are in areas of high biodiversity but little is known about Asian and especially African cities regarding urban flora. In Latin America the situation is better but worldwide no floral lists exist for the 50 most populated cities (Clemants 2002). In Jinan City, China, a methodological comparison found that gradient analysis from the urban center to the fringe gives better estimates of the urban flora than the traditional block-area analysis (Kong & Nakagoshi 2005). In Taipei, green areas have 164 tree species (few shared among sites) and large evergreen native species dominate. Larger parks have higher richness, more landscape fidelity to the original vegetation, and more rare and endemic species (Jim & Chen 2007). In Africa, the urban areas of the Nile Delta (Egypt) have vegetation that is mainly correlated with moisture, pH, fertility and texture gradients, but plants always occupy sites similar to their natural habitats (Shaltout & El-Sheikh 2002).
Latin America has a long history of scientific study of urban biota, particularly the plants and there are several recent studies from México, Peru, Brazil, Chile and Argentina.
In the city of Ensenada, Mexico, there are 161 species, 61% non-native (Garcillán et al. 2009). In Mexico City, trees are stressed from dry wind and unfavorable water flow caused by the pavement (Barradas 2000).
Brazil has the largest urban forest in world (Tijuca: 3 300 hectares) but it is being stressed by roads because roads are surrounded by invasive species that burn easily. The fires in turn open adjacent areas to more invasive vegetation and the damage spreads (Matos et al. 2002). There are very few studies of plants that grow on walls but in Jundiai, Brazilian, walls have a biodiversity of 28 species (dos Reis et al. 2006).
Normally, satellites are not used to study urban vegetation but in Arequipa, Peru, satellite images show that desert vegetation is being lost because of urban expansion (Polk et al. 2005). However, only ground work can reliably identify species and this kind of work has shown that temperate South America is not different from North America and Europe: at least half of the plant species in the Argentinean cities of Mendoza and Rosario are introduced. In Luján de Cuyo, Mendoza, 61 species were identified: 69% introduced (Méndez 2005). The vacant lots of Rosario each have one dominant species, a few abundant species and many rare species. Therophytes predominate and the proportions of indigenous and introduced species are similar (Franceschi 1996).
Chile is the Latin American country with the largest number of recent studies. Synanthropic communities in an urban footpath of Valdivia represents six associations and two communities (Finot & Ramírez 1998). In Concepción, green areas are dominated by nonnative ornamental species (Paucharda et al. 2006). The distribution of urban vegetation reflects social inequalities. In Santiago, poor areas can have ten times less plant cover than rich neighbourhoods (Hernández 2008), similar to other countries (Pedlowski et al. 2002). However, workshops in poor areas of cities can result in an improvement of their vegetation (Garzón et al. 2004).
In Costa Rica, there is a long history of study of urban plants that began with the National Museum's collection efforts in the late 19 th century, but little has been published. Méndez & Fournier (1980) and Monge-Nájera et al. (2002a,b) studied the lichens and their relationship with air pollution. The use of European lichens proved succesful when they were transplanted to this Tropical city (Grüninger & Monge-Nájera 1988). Francisco Fallas made checklists and abundance estimates of urban herbs in the late 1970's but to our knowledge he did not publish them. The program "Costa Rica: Jardín Botánico de América Tropical" produces manuals and labels for urban vegetation (www.hjimenez.org) and there is a program to provide urban parks with butterflies and their host plants (http://www.lrsarts.com/ plas/index.html).
The biodiversity in patches of urban vegetation can be surprinsingly high, at least in Costa Rica. For example, after 50 years, in only one hectare of urban vegetation in San José, there are 432 plant species (Di Stéfano et al. 1995, Nishida et al. 2009 We present an estimate of how the vegetation has changed in the city of San José, Costa Rica, after about one century, with a technique called repeat photography. We do not know of repeat photography studies on urban Costa Rican vegetation of San José, but the technique was used in Costa Rica by Horn (1989) to assess changes in the páramo habitat.

MATERIALS AND METHODS
We used a collection of photographs taken in the late 19 th and early 20 th centuries from the José Fidel Tristán Fernández Collection in the Archivo Nacional de Costa Rica and others reproduced by Leiva (2004). The sites were re-photographed on December 11, 2008 with a Nikon Coolpix 8800 camera (8 megapixels; We used digital repeat photography, which is fast, detailed and reliable; can be stored for future corroboration and comparison; includes rich data that may become useful in the future; and can classify and measure information automatically (Crimmins & Crimmins 2008).
We re-photographed nine sites ( We calculated % cover by clipping and weighing sections from photographs printed on standard bond paper; for example, if the clippings from buildings represented 20 % of the total weight of the photograph, we recorded that buildings represented 20% of the image. This technique is compared with others by Monge Nájera et al. (2002b).
We also present an analysis of plants collected in cantón de San José from 1885 through 1945 by the Museo Nacional staff (Base de Datos, Herbario Nacional, updated to April 13, 2009).

RESULTS
After about a century, the main change in San José city photographs is the much larger number of people. The reduction in vegetation affects grasses, shrubs and trees, but is small; the increase in buildings, streets, vehicles, sidewalks, lamps and signs is also small (Fig. 2).
A total of 95 families and 458 species were collected in the late 19th and early 20th century.

DISCUSSION
Repeat photography was used in Costa Rica by Horn (1989) to assess changes in the paramo habitat (she found very little change) but we were unable to find a similar study of urban vegetation. Nevertheless, studies about plant biodiversity after long periods suggest that our results are not unusual. For example, after 50 years, Brussels has the same total number of plant species that it had in 1940, albeit the individual species change and the same applies to other areas in much longer time spans (Chocholouškováa & Pyšek 2003). We cannot make a comparison of species from circa 1900 to circa 2000 in San José because urban vegetation has rarely been collected in recent decades.
The increase in human presence in the city landscape is explained by the population growth of downtown San José or Cantón Central (from about 30 000 when the first photographs used in this study were taken circa 1900 to 356 000 when the sites were rephotographed in 2009; see: Centro Centroamericano de Población and Instituto Nacional de Estadística y Censos 2002). The small increase in the number of vehicles is an underestimation: the 2008 photographs were purposefully taken in low traffic periods to obtain a better view of the scenes.
Central European cities have a mean of 646 plant species/city and larger cities have more species (Pysek 1998), thus, considering that San José was and is a small city, the total of 458 species recorded is within the expected range.
Successful urban plants tend to belong to species adapted to natural habitats with strong sunlight, abundant nitrogen and low    (Abdollahi et al. 2000, McPherson et al. 2008, significantly reduce urban heat islands (Huang et al. 2009) and can sequester about 100kg of air pollutants per hectare of urban forest (Vilela-Lozano 2004). Furthermore, urban vegetation protects many species in the five kingdoms (Dana et al. 2002, Smith et al. 2006, including valuable rare species (Maurer et al. 2000, Williams et al. 2009). For these and other reasons, the study and management of urban vegetation is of great importance.
Genetic diversity is low in urban plants and they are less prepared to cope with environmental change (Knapp et al. 2009), so periodic monitoring is needed to conserve original species as well as any others in need of protection (Godefroid 2001). Citizens can learn to effectively take advantage of urban vegetation (Garzón et al. 2004) and to recognize historical changes in the city scene (e.g. the repeat photography groups in www.flickr. com), not only for cultural reasons, but also to influence the administration of the urban flora by local governments.
Future studies of urban vegetation in San José could investigate these hypotheses: Floristic composition results from the inter- With global warming, species adapted to • colder climate will become less common and vice versa.

ACKNOWLEDGMENTS
We thank Hubert Blanco (Archivos Nacionales), Colegio Superior de Señoritas, María José Guerra Araus (Herbario Nacional), Sergio Aguilar, Sergio Quesada, Karla Vega, Andrea Sánchez and María Acuña for their assistance. We specially thank Sally P. Horn (University of Tennessee, Knoxville) for suggestions to improve the manuscript and for advice on repeat photography.