Sytze de Bruin

Assessing fitness for use: the expected value of spatial data sets

This paper proposes and illustrates a decision analytical approach to compare the value of alternative spatial data sets. In contrast to other work addressing value of information, its focus is on value of control. This is a useful concept when choosing the best data set for decision making under uncertainty due to error in the reported data. Application of the concept requires probabilistic accuracy measures and a loss function representing the cost of incorrect judgement about some target property. This is illustrated by an assessment of the suitability of two digital elevation models (DEMs) for determining the volume of sand required for building a container port. To demonstrate flexibility of the approach, accuracy assessment was based on both a random and a systematic sample of error data, using design-based estimation and model-based prediction, that is geostatistics. Analysis results included the expected loss for each combination of DEM and sampling strategy. These indicated that both DEMs were equally suitable for the intended use. Operational practicability of the method is highly dependent on the willingness of database producers to give access to sample information similar to the quick looks provided to potential users of remote sensing imagery.

A Bayesian approach to combine Landsat and ALOS PALSAR time series for near real-time deforestation detection

To address the need for timely information on newly deforested areas at medium resolution scale, we introduce a Bayesian approach to combine SAR and optical time series for near real-time deforestation detection. Once a new image of either of the input time series is available, the conditional probability of deforestation is computed using Bayesian updating, and deforestation events are indicated. Future observations are used to update the conditional probability of deforestation and, thus, to confirm or reject an indicated deforestation event. A proof of concept was demonstrated using Landsat NDVI and ALOS PALSAR time series acquired at an evergreen forest plantation in Fiji. We emulated a near real-time scenario and assessed the deforestation detection accuracies using three-monthly reference data covering the entire study site. Spatial and temporal accuracies for the fused Landsat-PALSAR case (overall accuracy = 87.4%; mean time lag of detected deforestation = 1.3 months) were consistently higher than those of the Landsat- and PALSAR-only cases. The improvement maintained even for increasing missing data in the Landsat time series.

Development of a Dynamic Web Mapping Service for Vegetation Productivity Using Earth Observation and in situ Sensors in a Sensor Web Based Approach

This paper describes the development of a sensor web based approach which combines earth observation and in situ sensor data to derive typical information offered by a dynamic web mapping service (WMS). A prototype has been developed which provides daily maps of vegetation productivity for the Netherlands with a spatial resolution of 250 m. Daily available MODIS surface reflectance products and meteorological parameters obtained through a Sensor Observation Service (SOS) were used as input for a vegetation productivity model. This paper presents the vegetation productivity model, the sensor data sources and the implementation of the automated processing facility. Finally, an evaluation is made of the opportunities and limitations of sensor web based approaches for the development of web services which combine both satellite and in situ sensor sources.

Regional and local maize seed exchange and replacement in the western highlands of Guatemala

Regional distributions of crop diversity are important to take into account for the spatial design of in situ, farmer-participatory interventions in crop genetic management. Regional seed flows are an important factor in shaping geographical distributions of crop diversity. This study contributes to the insight in these seed flows, focusing on maize (Zea mays L.) in Chimaltenango, an area in the western highlands of Guatemala. A social survey of 257 households on different aspects of seed management produced information on cultivar naming, seed sources, reasons and causes of the discontinuation of seed lots, and important explanatory variables associated with different seed sources. A small portion of the reported seed lots originated from regional seed sources. The main motivation of regional seed exchange and the discontinuation of seed lots was to achieve change in plant characteristics of the crop, especially to obtain lower plants and shorter growing cycles. It is argued that farmer selection fails to achieve such change, and in fact leads to an equilibrium with high plants and long growing cycles. Seed exchange functions as an escape to this trend. Other factors of influence on seed exchange are altitude and ethnicity. The study also highlights the issue of geographical directionality in seed exchange patterns.

Modelling Positional Uncertainty of Line Features by Accounting for Stochastic Deviations from Straight Line Segments

The assessment of positional uncertainty in line and area features is often based on uncertainty in the coordinates of their elementary vertices which are assumed to be connected by straight lines. Such an approach disregards uncertainty caused by sampling and approximation of a curvilinear feature by a sequence of straight line segments. In this article, a method is proposed that also allows for the latter type of uncertainty by modelling random rectangular deviations from the conventional straight line segments. Using the model on a dense network of sub-vertices, the contribution of uncertainty due to approximation is emphasised; the sampling effect can be assessed by applying it on a small set of randomly inserted sub-vertices. A case study demonstrates a feasible way of parameterisation based on assumptions of joint normal distributions for positional errors of the vertices and the rectangular deviations and a uniform distribution of missed sub-vertices along line segments. Depending on the magnitudes of the different sources of uncertainty, not accounting for potential deviations from straight line segments may drastically underestimate the positional uncertainty of line features.

Spatial Accuracy Assessment and Integration of Global Land Cover Datasets

Along with the creation of new maps, current efforts for improving global land cover (GLC) maps focus on integrating maps by accounting for their relative merits, e.g., agreement amongst maps or map accuracy. Such integration efforts may benefit from the use of multiple GLC reference datasets. Using available reference datasets, this study assesses spatial accuracy of recent GLC maps and compares methods for creating an improved land cover (LC) map. Spatial correspondence with reference dataset was modeled for Globcover-2009, Land Cover-CCI-2010, MODIS-2010 and Globeland30 maps for Africa. Using different scenarios concerning the used input data, five integration methods for an improved LC map were tested and cross-validated. Comparison of the spatial correspondences showed that the preferences for GLC maps varied spatially. Integration methods using both the GLC maps and reference data at their locations resulted in 4.5%–13% higher correspondence with the reference LC than any of the input GLC maps. An integrated LC map and LC class probability maps were computed using regression kriging, which produced the highest correspondence (76%). Our results demonstrate the added value of using reference datasets and geostatistics for improving GLC maps. This approach is useful as more GLC reference datasets are becoming publicly available and their reuse is being encouraged.

Updating cover type maps using sequential indicator simulation

Abstract Maximum posterior probability (MAP) maps of forest inventory (FI) cover type classes were produced from a maximum likelihood (ML) classified TM image and 5% (2%) systematic reference sampling of actual cover types for of nine 2×2 km study sites in New Brunswick, Canada. MAP cover type maps were obtained via sequential indicator simulation (SIS) using collocated indicator cokriging. A 5% reference sampling increased the coefficient of accuracy of MAP cover type maps by about 0.2 compared to the accuracy of the ML classified maps. MAP prediction errors were obtained for global and small area estimates of cover type extent. MAP-based cover type statistics of extent and precision were compatible with corresponding results for maximum likelihood bias-corrected estimates (MLE). Spatial autocorrelation of MAP prediction errors declined rapidly with distance and were near 0 for distances of more than 3–4 Landsat TM pixels. MAP cover type maps produced by SIS are attractive when both global and local estimates of precision of map-derived statistics are needed.

Design and Implementation of an Interactive Web-Based Near Real-Time Forest Monitoring System

This paper describes an interactive web-based near real-time (NRT) forest monitoring system using four levels of geographic information services: 1) the acquisition of continuous data streams from satellite and community-based monitoring using mobile devices, 2) NRT forest disturbance detection based on satellite time-series, 3) presentation of forest disturbance data through a web-based application and social media and 4) interaction of the satellite based disturbance alerts with the end-user communities to enhance the collection of ground data. The system is developed using open source technologies and has been implemented together with local experts in the UNESCO Kafa Biosphere Reserve, Ethiopia. The results show that the system is able to provide easy access to information on forest change and considerably improves the collection and storage of ground observation by local experts. Social media leads to higher levels of user interaction and noticeably improves communication among stakeholders. Finally, an evaluation of the system confirms the usability of the system in Ethiopia. The implemented system can provide a foundation for an operational forest monitoring system at the national level for REDD+ MRV applications.

Value of information and mobility constraints for sampling with mobile sensors

Wireless sensor networks (WSNs) play a vital role in environmental monitoring. Advances in mobile sensors offer new opportunities to improve phenomenon predictions by adapting spatial sampling to local variability. Two issues are relevant: which location should be sampled and which mobile sensor should move to do it? This paper proposes a form of adaptive sampling by mobile sensors according to the expected value of information (EVoI) and mobility constraints. EVoI allows decisions to be made about the location to observe. It minimises the expected costs of wrong predictions about a phenomenon using a spatially aggregated EVoI criterion. Mobility constraints allow decisions to be made about which sensor to move. A cost-distance criterion is used to minimise unwanted effects of sensor mobility on the WSN itself, such as energy depletion. We implemented our approach using a synthetic data set, representing a typical monitoring scenario with heterogeneous mobile sensors. To assess the method, it was compared with a random selection of sample locations. The results demonstrate that EVoI enables selecting the most informative locations, while mobility constraints provide the needed context for sensor selection. This paper therefore provides insights about how sensor mobility can be efficiently managed to improve knowledge about a monitored phenomenon.

Integrating global land cover datasets for deriving user-specific maps

ABSTRACT Global scale land cover (LC) mapping has interested many researchers over the last two decades as it is an input data source for various applications. Current global land cover (GLC) maps often do not meet the accuracy and thematic requirements of specific users. This study aimed to create an improved GLC map by integrating available GLC maps and reference datasets. We also address the thematic requirements of multiple users by demonstrating a concept of producing GLC maps with user-specific legends. We used a regression kriging method to integrate Globcover-2009, LC-CCI-2010, MODIS-2010 and Globeland30 maps and several publicly available GLC reference datasets. Overall correspondence of the integrated GLC map with reference LC was 80% based on 10-fold cross-validation using 24,681 sample sites. This is globally 10% and regionally 6–13% higher than the input map correspondences. Based on LC class presence probability maps, expected LC proportion maps at coarser resolution were created and used for characterizing mosaic classes for land system modelling and biodiversity assessments. Since more reference datasets are becoming freely accessible, GLC mapping can be further improved by using the pool of all available reference datasets. LC proportion information allow tuning LC products to specific user needs.