1Urban French Geography gave less importance to environmental issues and to nature/society relationship than rural geographers (Mathieu, 1992). It is rather recently that the topic of sustainable towns led to the analyse of environmental problems within metropolitan areas (Senecal, 1996). But those researches clarify neither what kind of links exist between nature in town and environmental problems, nor what are the geographical concepts to be mobilised for such an investigation. In our team, the idea of reintroducing nature and physical Geography in urban spatial analysis takes its origins from two roots.
2First a theoretical point of view (Blanc, Mathieu, 1996). Analysing environmental problems in towns requires a fundamental interrogation on the opposed concepts of nature/urban, natural "milieu"/ artificial/milieu. Why is it impossible to consider that there are natural phenomenon inside urban areas ? Why an urban "milieu", even in its "green" spaces, is always conceived as mastered by man and human technical power ? Why do urban geographers refuse the ecological examination of their field ? Our main assumption is that those oppositions can be overcome as well as the lack of interest in spatial urban analysis for the materiality of urban space by studying frontally the topic of nature in town and renewing with internal interdisciplinarity inside Geography between physical and human geographers both using spatial analysis (quantitative and qualitative tools).
3Secondly an interdisciplinary program between ecologists/ethologists and geographers investigating the man/animal relationship in two urban micro-areas (Mathieu et al., 1997 ; Blanc et al, 1999). The problem was the regulation and control of pullulating cockroaches (non-desired animal) in a social-housing district (South ZUP in Rennes) and epidemic risks linked to the multiplication of stray cats (desired animal felis catus) due to people feeding them near the Croix-Rousse hospital in Lyon. The results pointed to the need, when modelling a micro urban spatial system in which inhabitants and their qualitative life style are taken into account, to match two types of information, on the one hand the social representations of nature – what we call "urban-inhabitants cultures of nature"- ,and, on the other hand, the physical dimensions of urban places – what we call materiality of the urban nature in a town.
4The results of both investigations, fundamental and applied, were converging towards the necessity of reintroducing descriptive data on representations and practises of nature in urban areas as well as on real natural and environmental phenomenon. This paper focuses on the presentation of what sort of variables can fulfil our target. How can we describe "nature" in a town, "natures" of a town ? What is the difference between urban environmental criteria (sound pollution, atmospheric pollution, soil pollution…) and urban nature which is traditionally limited to "green space", gardens and parks, vegetation and trees ? How can we introduce quantitative data on animals in towns and objective measures of the man/animal relationship ? We try to resolve all those methodological problems through the well-known square-grid GIS of Rouen.
5The first stage of our work was to define "nature", frequently confused in common language with "environment". Then we set about identifying the different forms in which nature is present in an urban setting.
6The problem is the frequent confusion between "nature" and "environment"/ "environmental problems". The use of the term "environment" implies in fact a more or less moral qualification applied to natural phenomena in interaction with human societies. "Environmental problems" refers to interaction harmful fot the carrying out of human activity (floods, earthquakes, etc.), as well as to the upsetting of these natural phenomena provoked by human activity (global warming, pollution, etc.). As for environment, this corresponds to the materiality of the ideal of nature that urban societies have constructed for themselves. In fact, urban green belts or the preservation of natural havens (national nature reserves) correspond more to an idealisation of nature than to a reality.
7Actually, the environment, and more particularly in the urban setting, can be considered as the aesthetic and ethical side of nature.
8So what should we understand by "nature in the city", which is such a polysemic concept ?
9Finally, we have therefore defined as "nature" everything that resists, surprises, worries or fascinates us, all the causal determinations that we can study, which we can act upon but not suppress.
10From now on, in the urban setting, we must therefore consider as nature all that constitutes its materiality. To study and to look into nature in the city is also to study its physical materiality, its basic nature.
11Does this mean, then, that to study nature in the city is to study all of it ? Certainly not ; our aim is not to create an exhaustive catalogue but rather to question the modes of geographical analysis on the city by re-examining one of the basic paradigms of the discipline, that is the relations between society and nature. Therefore, we have focused our research only on those elements of nature, which are tangibly in interaction with urban populations, and recognised by them. Through these interactions, both "objective" (by objective we mean spatial, statistical and topological) and "subjective", we are seeking to establish a new definition of urban geographical milieus and to sketch what would be an urban culture of inhabiting.
12According to the degree of naturality or artificiality of natural elements existing in the urban milieu we may differentiate between several levels and several qualities of nature in the city:
-
Nature known as "wild", that is, not subject to any human influence. The elements which make this up may be either positive or negative for human habitation. Although it is of necessity conditioned by human activity in a world in which man occupies all the space, we can consider this nature to be "wild" as far as its existence and its localisation are not subject to the will or the "art" of humans. We may place in this category the setting, meteorological phenomena, vegetation, wild animals and watercourses.
-
Domesticated nature, that is elements that have been submitted to the use and wishes of humans. Although they are subject to human power they keep nevertheless a certain autonomy and an inherent character, which prevent them from being simply artefacts. These are mineral, animal or vegetal elements which societies have made their own in order to embellish (green parks, pets), build (the different buildings, tarmac) or protect themselves (river dams). This is nature in the service of man and his artefacts.
-
Nature as a reaction to an "excessive" development of artificiality. By essence, or rather by definition, human art lies outside the natural sphere and if it can sometimes dominate nature, there are also, necessarily, conflicts. These express themselves through damage, parasites, and other risks and natural catastrophes. This nature is made up of atmospheric and other pollutants, parasites (rats, cockroaches, termites, etc.), floods and so on. In a word, everything that gradually destroys or takes over artefacts and harms urban societies. That is hazards and risks of natural environment.
13The second stage is the GIS we are building in order to conduct an objective analysis of man/nature relations in an urban area.
14According to the preceding redefinition of "nature", we have identified and defined the variables and the data theoretically necessary to describe the natural environment. We distinguish :
15In the wild nature "category" :
-
The system of slopes which refers to the setting.
-
The climate (that is, temperature, wind direction, length of sunlight hours)
-
The axes of penetration by rural animal species in the urban area
-
Density of birds
-
Diversity of small mammal species
-
In the domesticated nature category :
-
Plant cover in private and public areas
-
Pets (dogs and cats)
-
In the "nature as parasite" category :
-
Animals [cockroaches and fleas, birds (seagulls, starlings and pigeons) and stray animals (dogs and cats)
-
Erosion of buildings
-
Car fume emissions
-
Dominant smells
16In addition, in order to make our work as complete as possible, we must also take into account what nature, by default, implies, that is, artefacts. We have thought it was necessary to take the following information into account :
-
Type of building, ie size, area, number of floors and building materials
-
Mineralised ground areas.
-
And, the sources of polluting emissions by industries
17Some variables are already known and collected as Yves Guermond’s laboratory (MTG, "Modélisation et traitements graphiques") is building a GIS on the urban area of Rouen since 10 years (Guermond, Lajoie, 1996). So, the slope system has already been integrated by using the DTM according to square grid cells of 75 metre length. We also have the localisation of mineralised ground surfaces, the sources of polluting industries and polluted soils but nothing about the other forms of nature in town.
18So, the diversity and the novelty of the information which we need mean that, most of the necessary variables and data will have to be created by our own research team. Although it has been possible to create some from detailed information.
19This is the case for the axes of penetration of rural species in the urban area, for polluting birds (the areas occupied by seagulls and starlings are very small and they are always located in the centre of the urban area. And, as populations are not so great, it’s quite easy to make an exhaustive census. The pigeons, being in great numbers, are more difficult to count, so we chose another way to collect the data.).
20We also created the plant cover layer by the analysis of a multispectral Spotimage with the NDVI. It was a summer image. So, this information about the plant cover must be completed. In fact, the form of vegetation in town depends on the season and, we don’t have any information about species by this way. Consequently, this layer is only a part of the reality of vegetation in the urban area.
21Moreover some data have had to be collected by means of sampling. Because of the large size of the zone under study, it has been necessary to work on reduced test zones ; this was the case for bird density, diversity of mammal species, smells from car pollution and climate. The difficulty is to define criteria, which allow an extrapolation on the scale of the whole urban area.
22Finally, some variables come from indirect information. This is the case for the inventory of stray animals and cockroaches and fleas infesting the home. To locate cockroaches, we take and locate within the road network the addresses of complaints made to municipal hygiene services. And, to locate stray animals, we geolocate the addresses where stray dogs and cats were caught by the stray animal catching service. The data that we have at our disposal, do not therefore correspond to the reality of their presence but rather to the level of non-tolerance of them by human populations. In fact, this information is more a part of the next stage of our work, that is, the study of objective relations between urban people and nature.
23Finally, these data make a problem because of their mode of spatial distribution. In fact, they were collected in very different spatial forms : selective when there is an address, surface-related for building and plant elements for example and fragmented when it is a question of poll results in a test zone. These latter should be able to be extrapolated for the whole urban area by following criteria of choice for the test zones (for example, for avian density, we have taken into account the amount of plant cover, the size of building and the distance from the city centre ; these criteria are recognised as determining factors by ornithologists to explain the variations in density of avian populations in the urban milieu. For example in the town centre there are more birds because more food is available and there are more places to nest).
24Then, to be able to make crossover comparisons between the different layers of information, it has been necessary to put in place a spatial division of reference into which we can integrate all the data.
25We therefore chose the square grid system of which the M.T.G. laboratory in Rouen already had a good deal of experience. We retained 250-metre-long cells, which enables to work on a large scale without excessively multiplying the number of squares. Thus, to cover the agglomerated zone of the Rouen district we already have 2298 squares (5178 to cover the suburban divisions). In addition, the size of these squares corresponds approximately to that of the smallest census id. defined by the INSEE (French national institute of economic and statistical information) for its general population census whose data we shall use to characterise human populations. Finally, on this point, the square grid system enables us to use these data per block without violating the CNIL ruling which forbids publication of information on zones with fewer than 2000 inhabitants.
26From a technical point of view, the data already collected and integrated in the GIS are entered in the square grid by spatial aggregation where selective information is concerned (each square contains in its data the number of items in its area). they are integrated by spatial re-division when they are surface related (each square then includes a certain percentage of its surface covered by a particular type of building or particular plant cover for example). Finally the criteria used for the polls are given in the squares in order to allow an extrapolation of its results on the scale of the urban area. In reality, most of the polls were carried out after the division into the square grid, so the test zones being defined by the squares, it was easier to make extrapolations.
27At this stage of the progress of our work, the more important layers of the GIS are completed. However, these variables do not have a definitive value. Given the fact that this is an exploratory project, it seems that it will be necessary to add further variables later on.
28We must now carry out the analysis of the information collected and make a crossover of the different layers with a view to proposing objective measures of man/nature relations. Actually, as yet, we have made little progress in this field and we have only developed simple indexes such as the number of square metres of vegetation per inhabitant or the percentage of vegetal cover in relation to the slopes. In order to be able to make a complete typology of the squares according to the different levels of man/nature relations, it seems necessary to develop subsequently more complex indexes. These should be able to account for all the elements present with more complex indexes or collections of indexes.
© CYBERGEO