Changing Landscapes - Landscape resource assessment

In future, scenario studies will more often be used as a basis for making political decisions. This project will contribute with many relevant data for such studies. Behaviour of rural populations and the development of the rural landscape have always been controlled both by landscape resources and political and economical conditions.

The key to sustainable land management is detailed landscape resource information, and in recent years there has been a sharp increase in the demand for such information. This is in response to the necessity of implementing the environmental policies of Danish government authorities--, on for example set-aside, nitrogen application and afforestation, which have been effective for some years.

Efficient implementation of future policies requires very detailed landscape information, since it is only possible to mitigate the negative environmental impact of anthropogenic perturbations with a detailed knowledge of the land- and nature resources.

This project aims to develop new methods to describe the landscape resources (soil, climate and topography) on a detailed scale. Geostatistical methods are applied in order to develop soil classification systems and to quantify soil spatial variability.

The quaternary geological map from Geological Survey of Denmark and Greenland (GEUS) provides unique information on parent material, which is a key parameter in soil classification. Analyses of these maps combined with the soil databases from Dept. of Agricultural Systems (JBS), would provide a new set of soil data on a national level.

The data will provide a basis for environmental risk assessment, forestry research, and physical landscape planning and bottom-up modelling of run-off, leaching of N, P and pesticides.

Background

Danish government authority policies, on for example set-aside, nitrogen application and afforestation, have been effective for some years. Efficient implementation of such policies requires very detailed landscape information, since it is only possible to mitigate the negative environmental impact of anthropogenic perturbations with detailed knowledge of the land resources.

In recent years there has been an increase in the demand for detailed landscape information, which is essential if the land is to be managed in a sustainable manner. Foresters need the information in planning afforestation, for example to choose right tree species, avoid frequent wind throw and severe forest die-back. Farmers need land information when fertilizers and pesticide are to be applied according to soil type on the individual field. When modelling the risk of erosion and nitrate and pesticide leaching to the ground water, information on soil, climate and topography is essential. Use of very detailed landscape information gives the possibility of optimizing land use in both the environmental and economic sense.

In order to predict the environmental consequences of future changes in land use scenarios must be investigated. Maps of soil erosion from the different scenarios can be compared and used to indicate where possible control measures would have the greatest positive and negative consequences. Planners can then decide how to combine the positive elements from several scenarios.

The Great Danish Land Register of 1844 was the last detailed land assessment. This assessment was carried out because there was an urgent need for an improved basis for national taxation. The main goal of a new land assessment is to produce the basis for a sustainable management of the land resources. At present there is neither up to date soil information available at the field scale nor any applicable methods available for carrying out a detailed land resource assessment in Denmark.

Currently there is a debate emerging amongst soil scientists regarding conventional and continuous methods of soil classification. Conventional classification methods establish a series of subdivisions which place individual soil profile descriptions into hierarchical schemes. The schemes prevalent in many countries, methodically constructed and developed during the past half century, rely on a modal soil profile to present a land unit. The precision of any prediction using this classification is dependent on the homogeneity of the mapping unit and on the spatial variability within mapping units. The spatial variability is typically not ascertained, and without a measure of the spatial variability within each mapping unit, little is known regarding the reliability of the modal soil profile. Continuous classification is possible by using different types of geostatistical tools, which are capable of accounting for the continuous nature of soil and allows an individual soil profile to be assigned totally, partially or not at all to a particular class.

The land evaluation system termed "Forstlig Lokalitetskortlægning" has been used extensively by the National Forest and Nature Agency to map afforestation land since 1990.

Evaluation of available water capacity, soil nutrient status and factors limiting to root growth are based on judgement in the field and in turn used for recommendation of suitable tree species. Regarding soil nutrient status there is, however, a need for base-line information to form a quantitative scale. The long-term nutrient regime of soils is a function of parent material, climate, topography and geological age, determining the potential weathering rate. Biology, former and present land-use (human impact) may have altered this either by amelioration or degradation of the original potential. Knowledge about soil fertility is of great importance in agriculture e.g. for fertilisation planning and of special relevance in forestry, since amelioration measures and fertilisation are not common. The evaluation of long-term carrying capacity of soils for different land-uses should ideally be based on estimated soil fertility expressed as soil nutrient status.

Deriving spatially distributed estimates of meteorological data are becoming increasingly important as inputs to various models. The techniques for generating these input data are continuously being investigated. Recently, the Danish Meteorological Institute has begun releasing gridded precipitation data using a kriging interpolation method. Attempts to map areas exposed to frost and droughts call for calibration and investigation of existing techniques for other meteorologi-cal parameters. Especially when using existing water balance models to estimate the risk of drought in afforestation schemes there is a need for input of several types of data (meteorology, topography, vegetation and soil). Furthermore there is a need to calibrate these models to different types of forest.

Project objective

The main objective of the project is to develop a method for landscape resource assessment in order to facilitate sustainable land use and environmental risk assessments.

Research objectives

This project aims to develop methods describing the landscape resources (soil, climate and topography) on a detailed scale. These data will be an excellent basis for bottom-up modelling on run-off, loss of N, P and pesticides.

· Develop a method for mapping the Danish soil resources at large scale (1:10000), and establish the framework for a Danish system of soil series.

· Quantify and compare the complexity of the soil pattern in different geological environments.

· Utilise the quaternary geological map to soil classification on a national scale.

· Quantify the quality of the chosen mapping method.

· Develop a method for identifying areas susceptible to frost and drought.

· Develop a pedo-transfer function for soil nutrient status.

· Analyse and extract all useful soil information from a digital version of The Great Danish Land Register of 1844.

The project form the detailed physiographic basis for:

Physical landscape planning.
Planning of new land use.
Landscape analysis.
Sustainable agricultural and forestry production.
Assessing nature quality.
Analysing landscape structure and functionality.
Analysing economic consequences of authority measures in the landscape.
Analyse probability of human settelment and landscape history on a long- term scale.

Work content

Research strategy

1. Establish a revised Danish texture classification system on the basis of the existing Danish soil database (JBS database).

2. Establish the framework for a Danish soil series system on the basis of the JBS database.

3. Develop a method for a detailed survey (1:10000). Carry out a detailed soil survey of the project area(s).

4. Quantify soil variability in relation to mapping units, landform, topography and parent material.

5. Develop a method to quantify the quality and purity of the map units

6. Establish a new national soil data set derived from a statistical analysis of the existing databases from GEUS and JBS.

7. Develop a model for providing a spatially distributed meteorological data set.

8. Develop a water balance model for young forest vegetation.

9. Quantification of soil nutrient status and linkage of nutrient status to key variables discernible in the field in a transfer function applicable in the general mapping system for Danish soils developed in the project.

10. Analyse the use of digital version of The Great Danish Land Register of 1844 in modern land assessment.

Methods

ad 1

All kinds of field investigations and classification of soils are dependent on a sensible and well documented texture classification system. The existing classification has some disadvantages which must be addressed in the new classification system.

The basis for this revision of the Danish texture classification system is existing data in the JBS soil database. A new classification must meet several conditions.

· One class must congregate soil with the same properties concerning, e.g. plant available water, risk of compaction and workability.

· One class must congregate soils derived from the same parent material.

· It must be possible to separate the different classes in the field.

The proposed textural classes must be tested against key soil parameters such as plant available water, retention capability of various chemicals and workability. In addition, it will be ascertained as to whether a skilled soil scientist can separate the classes in the field, with acceptable accuracy. Different statistical methods from the field of multivariate analysis can be applied. Cluster analysis can be used for grouping the multivariate observations into classes, and discriminant analysis can be used for classifying new observations onto the classes.

ad 2

The basis for carrying out soil survey is a field classification system. The characteristics used to differentiate a class in the classification system is parent material, lime content, texture, internal drainage, content of humus and the thickness and type of pedological horizons. The basic class in the classification system is the soil series. Initially the existing soil database on JBS is analysed and the knowledge and experience accumulated in the forest site classification is utilized. These data will be the basis for establishing the framework of a Danish system of soil series. As the mapping proceeds the system will be improved. Unknown combinations of diagnostic properties emerge and new soil series will be defined. Methods similar to this is currently used both in USA and in Canada.

ad 3

Reliable soil surveys can be made at reasonable costs because the locations of soils to some extent are predictable within the landscape. This soil-landscape relationship must be the basis for the method to be developed. In order to quantify soil variability and homogeneity in different areas the conventional soil survey is supplemented with geo-referenced drillings and the information from the drilling (for example horizon designation, thickness, colour, texture etc) are stored in a database. In this manner it is possible to carry out classification by using a decision support system, based on, e.g. Bayesian networks, fuzzy logic or neural networks. The method to be developed is a combination of the two traditional methods, free soil surveys and the grid method. When the soil survey is completed, a digital soil and terrain database is established and basic soil maps are produced.

ad 4 & 5

Little has been done in order to quantify spatial variability of soils in relation to landform under Danish conditions. Therefore the existing information from the JBS soil profile database will be extracted and analysed on landforms, in particular with a view to soil spatial variability. The existing information will be supplemented with data derived on the basis of test grids or transect in the project areas. Geo-statistics, consisting of variogram analysis and kriging will be applied as a tool for characterizing and mapping the spatial pattern of soil variables. Semivariograms are used to characterize and model the spatial variability of data, while kriging uses the modelled semivariograms to estimate values between sampled points by local weighted averaging. The method also allows for designing the optimal sampling procedures for the main field work.

These methods also quantify the quality of the prediction of the variables in given points. In the current situation the qualitative variable of soil series has to be handled.

ad 6

The systematic Quaternary geological mapping is a important data base, valuable in land use scenario studies. An untested application of this data is soil classification. Parent material is one of the key factors controlling soil formation. Therefore, the geological maps show soil texture and lithological units are important for the understanding of soil types distribution.

By now systematic Quaternary geological maps in scale 1:25.000 cover about 80% of the Danish area. These maps are now available on digital GIS-format and hence available for statistical analyses.

The aim of this project is to utilise these geological maps in soil classification based on statistical analyses of these maps combined with other types of existing digital, such as:

1. Pre-Quaternary surface.

2. Isopachyt map of Quaternary sediments.

3. Pre-Quaternary lithology map.

4. Areas with shallow groundwater and former wetlands.

5. Pleistocene - Holocene map.

6. Quaternary geological maps.

7. Danish Soil Profile Database

8. Data from the Danish Soil Classification.

9. Digital elevation model.

This study will be based on statistical analyses of the parameters that affect the soil formation and focus on extraction of as much information as possible from these data.

The output from the statistical analyses will be a GIS map in a quadratic grid (the size of the grid will depend on the data but at first a 1000 m * 1000 m grid will be used) with the calculated soil properties. The soil properties will be represented by average values from the soil types within the grid-cell (weighted by their area repre-sentation) together with expected range of values within the grid-cell calculated from soil properties within the soil type in the type area(s).

ad 7

Cold air regions are known to have strong correlation with terrain elevation, form and soil type. Utilisation of this relationship is possible within the framework of Geostatistical interpolation. The technique, known as co-kriging, allows a fine resolution map of frost prone regions to be produced on the basis of a detailed digital elevation data set. Combining this data set with detailed information of the spatial distribution of soil types with low thermal conductivity gives a unique possibility to locate areas exposed to early summer frost. Calibration of this model requires measurements of temperature with high spatial resolution when the meteorological conditions producing frost risk occur. Mobile surveys can provide these kinds of measurements at low costs. Problems with a time lapse can be overcome by having a reference measuring station in the project areas and conducting the surveys at night in clear, cloudless and calm conditions. If the project areas are small, it may be feasible to establish a network of small meteorological stations in the area thus providing measurements with both high spatial and temporal resolution.

ad 8

To evaluate the risk of drought on a large scale it is important to have information on soil types. In addition, the variation of meteorological variables, especially precipitation and incident global radiation with the latter corrected to ground truth is needed. Water balance models exist for calculating the risk of drought in agricultural crops. However, in these models parameters are needed for different types of forest. These parameters can be derived from the soil survey within this project, from other projects and the literature. Given input parameters of high spatial resolution it is then possible to estimate the risk of drought for different types of vegetation.

ad 9

By statistical analysis (e.g. multivariate clusteranalysis) of forest soil data from the DIAS database and the Nordsoil database (DFLRI) a scale for inherited soil nutrient status will be developed. These classes are tested by sampling of identified soil units in mapped project areas and permanent DFLRI trials. Soil samples from all genetic horizons will be described (field identifiable features) and quantitatively determined by use of conventional soil chemistry procedures and a weatherability index based on sequential 0.1M HNO3 extraction. The efficiency of conventional chemical methods and the weatherability test to characterise the soil nutrient regime will be evaluated.

ad 10

The Great Danish Land Register of 1844 is the only nation-wide detailed land assessments performed in Denmark. These old maps will be digitised and evaluated by comparing them to the modern soil survey carried out in this project.

Project area

The majority of the field investigations in this project will take place in the same areas. In this manner it will be possible to provide a total detailed description of soil, topography and climate.

This project has the following requirements for the project area:

High diversities in soil types and parent material.
High diversity in landforms.
Representing typical Danish landforms.
Representative meteorological station.
Inland location.

At present Bjerringbro/Hvorslev is chosen as one of 3 areas, the selection of the other 2 areas will be done in corporation with the other project under Changing Landscapes.

Digital elevation model

Essential to the more analytical part of the activities is establishing a detailed digital elevation model (DEM) for the project area. In connection with land resource assessment DEM´s have important applications:

Modelling surface runoff.
Visualising and classification of terrain (e.g. slope, slope length, aspect).
Correction of the meteorological model.

Simulating a runoff pattern in an area and calculation of slope, aspect and size and shape of the different catchments is essential for high quality soil erosion risk assessment, and also important in order to establish relationships between landscape and soil types.