Landscape functional mosaics
B�r�ard Meyer
1, Gábor Mezősi
21University of �ort�und, �epart�ent of Landscape Ecology
2University of Szeged, �eptart�ent of Physical Geography and Geoinfor�atics e-�ail: burhard.�eyer@ufz.de, �ezosi@geography.hu
________________________________________________________________________________
key words: landscape function, landscape �etrics, landscape functional �osaics
Introduction
Every landscape has its own potential, which has to be deter�ined in case of any kind of spatial planning �eant to be environ�entally friendly. �nowing the potential is also i�portant during the evaluation of the effective or opti�al land use. �ny other way, planning and research will not produce realistic results. Nu�erous �ethods and tools can be applied for the evaluation of landscape structure and processes. �hough, basically two approaches are accepted. On the one hand researchers �ight concentrate on spatial relationships and refer to the potentials of the landscape (Bastian, Schreiber 1999). On the other, if the operation and use of a given spatial unit are considered, one �ight end up in a functional landscape analysis. �he functions are not only physical and natural categories, but they are able to also signify a certain type of potential. �ost frequently landscape functions are defined as the pendants of landscape structure and landscape potential with an e�phasis on functional and social relationships. �o be aware of the landscape functions is crucial during �anage�ent, thus planners
�ust reveal these relationships if a well based pro��ect is expected. Besides, landscape functions have got an ele�entary role in deter�ining environ�ental risk, too (�ezősi, �lbrecht 2002).
�he �ethods used for the spatial analysis of the landscape are well docu�ented; the question of scale and structure is broadly discussed. Nevertheless, the spatial extension and hierarchy of landscape functions is so�ehow neglected in the literature. �his is understandable, since the evaluation of functions as spatial
�osaics raises a nu�ber of theoretical and practical questions. So�e of these are as follows. How can be an integrated/aggregate value understood fro� the aspect of the basic functions? What are the li�its of the theore� clai�ing that the analysis of functional �osaics �ust provide additional infor�ation on landscape quality and interactions?
�he �ain goal of this paper is to explore the integrated value of functional �osaics; to evaluate their suitability for deter�ining risk; and to asses the applicability of landscape �etrics in the description of landscape functional
�osaics.
Landscape functional mosaics
Functions in the landscape�ost of the interpretations agree that landscape functions can be divided into three the�atic groups (deGroot 1992, Bastian, Schreiber 1999, Bastian, Steinhard 2002). �hese are: econo�ic (e.g. renewable resources), ecological (e.g. resistance to soil erosion) and social (e.g. recreation). �he quantitative description of the nearly Klasyfikacja krajobrazu. Teoria i praktyka. Problemy Ekologii Krajobrazu. 2008, t. XX. 135-141.
three dozen of known landscape functions is not at all precise. �uring research, the �ost proble� is caused by scale. Landscape functions can be grouped at least into two classes of scale, thus if the above �entioned the�atic groups are considered too, then after all 1st order, 2nd order, 3rd order, �ain and subfunctions can be separated. If �ore functions are analysed during a research, it is very i�portant that they are at the sa�e hierarchical level. � further proble� is that functions, having already an aggregate character (e.g. the erosional and recreational functions are deter�ined by nu�erous factors), �ight show �ultiple co�binations at al�ost each spatial units, and �ight �obilise various relationships and interferences.
Analysis of landscape mosaics
�he �a��or concepts of landscape ecology are based on the assu�ption that the patterns perceived in the landscape are in close relationship with ecological processes (�urner 1989). �he �osaic character of the landscape and its features are well studied. It is also broadly accepted that the �osaic pattern is apparent regardless of scale (For�an 1995). Heterogeneity in the landscape appears in two ways. �he first is the gradient change of an attribute in space. Gradients do not have borders and no patches can be realised in the�, however they still represent heterogeneity (e.g. tropical rainforests, where changes can be considered continuous). Landscapes of this type are quite rare. �he second for� of spatial heterogeneity is represented by
�osaic patterns. In this case attributes are spatially organised, they for� aggregates, which can be delineated.
Landscape �osaics �ay contain patches and corridors. �his concept had set the base of the widely known patch-corridor-�atrix �odel (For�an, Godron 1986) and lead to nu�erical description of patterns with the help of index-lists. Patches can be influenced by the heterogeneity of substrates, natural disturbances and anthropogenic effects. Landscape co�ponents and attributes can be arranged into a patch-corridor-�atrix configuration.
�he evaluation of the landscape, including statistical analyses too, is highly dependant on scale. Landscape patterns are studied at three levels: patch-level, class-level (patch type level) and landscape-level. Landscape indices are derived fro� the para�eters defined for patches and patch types. Para�eters are usually su��ed or averaged, though the algorhyth� is frequently �odified by different authors. Landscape indices can be interpreted not only as si�ple heterogeneity indicators, and class indices also represent �ore than ��ust the frag�entation of the landscape. �hey are the �easures of landscape pattern as a whole.
the model of landscape functional mosaics
If landscape �osaics are defined as the aggregation of spatial patterns, landscape functional �osaics are then the aggregation of functional patterns, arranged in integrated spatial units. �ifferent landscape functional
�osaics can be created by focusing on different landscape functions during �odelling. In contrary to the �ost approaches in the last decade, the ter� landscape �etrics will be used in this study not only for land use or land cover. We apply landscape �etrics for the description of spatial data and for �apping/assessing landscape functions. �his results different levels of landscape �etrics on the sa�e cultural landscape.
�he application of landscape �etrics raises three �a��or proble�s. �he first is the integrated �anner of functions, and as a consequence their interdependency. �he second is the difference of scale. �he content and extension of a so called �ain function, such as pedological function is very different fro� that of a subfunction, e.g.
resistance to erosion. �he third proble� is the transfer of results (derived fro� the patch analyses of functional
�osaics) to the practise of planning and �anage�ent, and thus the way how functional values can be fitted to the existing landscape categories.
Met�ods
Landscape functions are scale dependent. �he functional difference at the existing hierarchy levels can be deter�ined, however, in order to exclude the serious proble�s originating fro� interdependency we applied ho�ogenous categories. For the analysis the following landscape functions, representing the third hierarchical
level, were selected: recreation, resistance to erosion, resistance to underground water and bio�ass production.
�he selection was �ade on the basis that these functions are the �ost i�portant in deter�ining environ�ental risk on the study areas.
Landscape functional �osaics were studied on a class-level. � �a��or proble� in this respect was the interpretation of the s�allest geo�etry where landscape functional ele�ents could be investigated (�ini�u�
ele�ent size was 500 �² and defined at a scale of 1:10.000). �he �ethodological analysis was focusing on two test sites, each with an area of 50 - 100 k�², situated in Ger�any and Hungary. �he Ger�an site, Jesewitz, is characterized by a �oraine landscape, and located north of Leipzig/Saxony. �he Hungarian site is a part of the Lake Velence catch�ent and includes the lake itself. �he surroundings of Jesewitz is characterised pri�arily by agricultural land use arable lands can be considered as the �ain land cover category. On the other hand, the Hungarian test site is characterised by co�bined land use and land cover (vineyard, recreation, agriculture).
Landscape �etrics were applied on the patches of the above deter�ined four landscape functions. Landscape potentials, described by �arks et al. (1992) gave the structural base of functions. Spatial units, fitting to the existing geotypes, were created this way. For exa�ple the resistance to erosion function is based on the resistance potential, which is dependant on the following factors:
- relief (the energy of overland flow is deter�ined by slope length and angle), - soil (physical soil type, hu�us-content, cobble-content),
- land use.
When calculating erosion probability, �arks et al. (1992) considered the sa�e factors, and used a table for�
evaluation. In our study �aps of physical soil types (9 categories), slope angle (11 categories) and land use (15 categories) were overlaid and intersected, in order to deter�ine the basic units (approxi�ately 800, representing nearly 70 classes) for which the likelihood of soil erosion was calculated. �he EPIC �ethod was used for this purpose, though Erosion 3� or any other Wisch�eier-S�ith based �ethod could also be suitable. Units were evaluated one by one, and e.g. where erosion probability proved to be very slight the soil resistance function was considered uni�portant. Patches were classified into 3 categories, and the spatial pattern of e.g. the soil resistance function was deter�ined this way. �hen the patterns were analysed on a patch-level, too, with the following �etrics: Shape Index, Peri�eter-�rea ratio, �ggregation Index, Core �rea Index, Proxi�ity Index.
However, several proble�s have turned up during the aggregation at this hierarchy level na�ely the nu�ber of functions was highly influenced by the geo�etry of the patches. �hus landscape �etrics were �ade only in case of the �ain landscape functions, otherwise several other functions should have been �athe�atically and statistically introduced. �nother �a��or question, worthwhile for further analysis, is how the interdependence between functions could be assessed and characterised.
First res�l�ts
�he data gained fro� the three base �aps (fig. 1, 2, 3, 4) were organized in a database with a 160 row x 180 colu�n setup. �fter overlaying the cell values of the different �aps we have calculated the functional characteristics (e.g. likelihood for erosion), which �ay support decision �aking in the future. �he patch values show that functions are usually of low spatial co�plexity (low peri�eter-area ratio), they are diverse (low Shannon’s index) and frag�ented (tab. 1, tab. 2).
It is hard to find a direct or visible relationship if the received �etric values are co�pared to the patch-level values of ecological factors (e.g. the physical characteristics of soils). Nevertheless, the a�ount of available data is not enough for a statistical evaluation. �hus, for further analyses a larger pool of data would be necessary.
�he question finally is: what additional �eans can landscape �etrics provide for deter�ining landscape functions and landscape potential, and how can changes be predicted in a cultural landscape? �his predo�inantly depends on the possibility of pro��ecting functional changes to the future. Regarding the Hungarian study area the extension of the nature protection function can be expected, which will decrease the weight of other functions (e.g. soil resistance), however, in the long run it �ay result an increase in the value of the soil productivity function. �his �ay also �ean that for instance in the case of the soil resistance function the �easured Shannon index or the Peri�eter-�rea Ratio can decrease, referring to an increasing ecological stability.
Fig. 2. Slope angle categories on the test area Fig. 1. Physical soils types
Fig. 3. Landuse �ap fro� 1999
Fig. 4. Resistance to erosion on the study site (� - �est �rea Jesewitz, Ger�any, B - �est �rea Lake Velence, Hungary)
A B
�able1. Calculated data for deter�ining the functional pattern of the �est �rea Jesewitz, Ger�any
Function Biopoly Boden Retention
�rea
Nu�ber of parcels 1260 1407 6326
�rea (Class) 66461099,97 61149497,14 61270744,02
Mean Patc� Size 251327,68 409011,74 100858,37
Patch Size stand. dev. 310474,94 988204,26 3419756,9
Edge Total Edge 1013427,23 1596793,27 2519756,9
Mean Patc� Edge 4900,62 10485,47 3919,37
�iversity
Richenes 8 10 11
Rel. richeness (%) 100 100 100
Shannon’s �iversity 0,683 1,72 1,721
Shannon’s evenness 0,328 0,747 0,718
�o�inance 1,397 0,583 0,677
Nu�ber of classes (act.) 8 10 11
Nu�ber of classes (pot.) 8 10 11
Acknol�edgament. Contribution financially supported by O��� fund �46558
References
Bastian O., Steinhard U., 2002. �evelop�ent and perspectives of Landscape ecology. �luwer, �ordrecht – Boston, 344.
Bastian O., Schreiber �. F., 1999. �nalyse und ökologische Bewertung der Landschaft. Spektru�, Heidelberg – Berlin, 564.
deGroot R., 1992. Functions of Nature. Wolters – Noorfhoff, 292.
For�an R. �. �., Godron �., 1986. Landscape Ecology. John Wiley & Sons, New York. 619.
For�an R.�.�., 1995. Land �osaics: �he Ecology of Landscapes and Regions. Ca�bridge University Press, Ca�bridge, 523.
�arks R., �üllers �. (ed), 1992. �nleitung zur Bewertung des Leistungver�ögens des Landscheftshaushaltes.
Zentralausschuss für deutsche Landeskunde, �rier, 222.
�urner �. G., 1989. Landscape ecology: the effect of pattern on process. �nn. Rev. Ecol. Syst. 20. 171-197.
�able 2. Calculated data for deter�ining the functional pattern of the �est �rea Lake Velence, Hungary �YPE�RE�
�RE�_ CPS
PERI�PERI�_ CPSP�R�
P�R�_ CPS
SH�PESH�PE_ CPSCORE
CORE _CPS
PROX
PROX_ CPS
161.091.30438200.091.304134.42624.34782.562591.30430.00.02.888952.174 194.095.652213800.0100.000146.80858.69573.4500100.00000.00.01.751043.478 1146.0100.000013400.095.65291.78080.00002.680095.65220.00.01.519239.130 2377.097.619031600.0100.00083.81962.38104.0513100.00000.00.02.113426.190 2433.0100.000021600.097.61949.88450.00002.571497.61900.00.02.598933.333 33277.0100.0000142200.0100.00043.39330.00006.1826100.00000.00.015.87417.391
