Since 2015 the EyeOnWater app and website aim to involve citizens in the observation of the colour of natural waters. In the App, the colour analysis is based on the Forel-Ule colour comparator scale. At the server a retrieval of the non-discrete hue angle (α) from the images by the WACODI algorithm supports the analysis. Now, over 5000 observations have been gathered on a global scale, including open ocean, coastal and inland waters. First, we present an introduction to the App and data handling. Second, we present an analysis of the data availability, regional and temporal distribution plus a number of worldwide hotspots where communities have taken up this monitoring tool.

Third, we will highlight the options to complement these Citizens observations by a number of past and present Space sensors. In 2016 we demonstrated that colour, expressed mainly by the hue angle (α), can be derived accurately and consistently from SeaWiFS, MODIS, MERIS and OLCI data. The algorithm consists of a weighted linear sum of the remote sensing reflectance in all visual bands plus a correction term for the specific band-setting of each instrument. These algorithms are operational and available in SNAP software. In 2018 the algorithm was extended to sensors like MSI on Sentinel-2, OLI on Landsat-8, Landsat 7 ETM+ and MODIS 500m products. The new products will be demonstrated for a number of Sentinel-3, Sentinel-2 and Landsat-8 images of complex inland waters.

Colour is a simple parameter that can be understood by the public in a fast and intuitive way. Adopting the same colour as a standard product from space missions will create a bridge of understanding between the space agencies and Citizen scientists.

Recent years have witnessed the practical value of open-access Earth observation data in land cover and land resources mapping. Combined with related free and open-source software (FOSS), cloud-based computing, and data collection through the crowd, possibilities for monitoring changes in land resources have improved substantially, especially in areas with difficult accessibility and data scarcity, and in the societies where access to data and technical capacity have been limited. To empower rapidly developing societies, there is increasing need to train professionals who master the use of these new technologies, methods and data. This development can be supported by introducing the open data, open source software, and open access cloud computing resources in the local higher education institutions (HEIs) remote sensing training.

In this study, we developed and tested a participatory mapping methodology utilizing visual interpretation of satellite imagery and cloud computing to map the previously unknown extent and species composition of forest plantations in the Southern Highlands area of Tanzania, a region experiencing a rapid growth of smallholder-owned woodlots. Our methodology utilizes open source software Open Foris and QGIS, various open Earth observation data catalogues, and computing capacity of the free Google Earth Engine cloud platform. We arranged a collaborative data collection event, Mapathon, in Tanzania, followed by a training of the related FOSS tools for local HEIs’ teaching staff. During the Mapathon, a team of 22 Tanzanian experts collected a large set of reference data focusing on forest plantations, by interpreting very high-resolution satellite images in Google Earth with the Open Foris Collect Earth tool developed by the Food and Agriculture Organization of the United Nations (FAO). We used the collected sample as training data to classify a multi-sensor image stack of Landsat 8 (2013-2015), Sentinel-2 (2015-2016), Sentinel-1 (2015), and SRTM derived elevation and slope data into a 30m resolution forest plantation map for our study area in Google Earth Engine.

The results show that the forest plantation area was classified with high overall estimated accuracy (85%), and the interpretation accuracy of local experts was high considering the challenges in general definition of forest plantations in the study area. Our results showcase the unique value of local expert participation, enabling the collection of thousands of reference samples over a large geographical area in a short period of time simultaneously building the capacity of the experts. This is evidenced though the feedback collected from the Mapathon participants about their learning experiences and the local HEIs teachers’ feedback about the usability of the methodology in remote sensing training in Tanzania.

Based on our experiences and the received feedback, using a participatory mapping campaign as a training method can offer effective learning about new technologies and tools in environmental remote sensing through a real-world example, as well as networking and knowledge sharing possibilities for the participating group. Since the methodology is built on open-access data and free software, it also presents a highly feasible solution for repetitive land resource mapping applicable at different spatial scales globally.

Keywords: Participatory monitoring, community-based monitoring, deforestation, Landsat, Sentinel-1, REDD+.

While large-scale forest monitoring systems typically rely on the analysis of remote sensing data, numerous case studies have also demonstrated that communities can play a pivotal role in the monitoring of forests. Thanks to their proximity and familiarity with the environment under concern, communities have the capacity to identify changes at an early stage, including small-scale disturbances and progressive forest degradation that remain for a large part undetected by remote sensing observations. The integration of participatory data streams in a national forest monitoring system – a key component of the REDD+ mechanism – has therefore the potential to enhance data exchange and increase transparency and stakeholder engagement for sustainable and climate-smart land use practices. This integration presents however socio-cultural and technical challenges that need to be addressed to provide mutual benefits and ensure the sustainability of the scheme.

Since 2014, Wageningen University and Research (WUR) has been working successively in Ethiopia, Peru and Indonesia to study the potential of interactive Near-Real Time (NRT) forest monitoring systems that combine remote sensing deforestation alerts with community-based monitoring (CBM) data collected using mobile phones. Following a first research project on CBM in Vietnam [1], an interactive web-based NRT forest monitoring system was implemented and evaluated in the UNESCO Kafa Biosphere Reserve in Ethiopia. Trained local experts could provide up-to-date and accurate information on forest change through mobile phones, that had the potential to be integrated into a national Monitoring, Reporting and Verification (MRV) system [2, 3]. Data collected by local forest experts over a 3-year period and dense-satellite time series spectral-temporal metrics were also used to train random forest models and predict deforestation, degradation and stable forests [4].

In Peru, the Ministry of Environment (MINAM) recently developed an online forest monitoring system providing deforestation alerts through the analysis of optical medium- and high-resolution imagery [5]. In addition, MINAM implemented Direct Conditional Transfers (TDC), a benefits scheme aimed at native communities for the conservation of forests. WUR supported MINAM by experimenting in 2016 the use of mobile devices for forest change data collection by 3 communities. In addition, 1,720 field observations collected manually by 43 communities from 2015 to 2017 were analysed to better understand challenges of the existing participatory monitoring system (incl. responding to satellite-based alerts) and evaluate how new technologies such as mobile phones or drones could empower the communities and contribute to a more efficient forest conservation scheme.

In Indonesia, WUR has been exploring since mid-2018 the development of an augmented interactive monitoring system relying on a NRT forest cover loss alerting approach based on Sentinel-1 radar data [6]. Radar-based forest cover loss alerts have the potential to overcome challenges met by optical-based systems, such as permanent cloud cover. Implemented in Google Earth Engine, a first prototype of alerting system has already been experimented on Flores Island in the scope of the LandSense project [7]. Further work will be conducted in 2019 to integrate participatory monitoring data streams and apply the a large-area interactive monitoring system for sustainable supply chain management in the oil palm industry.

We will present results of these 3 case studies and showcase the current progress of the new alert-driven interactive forest monitoring system based on Google Earth Engine. Our ultimate objective is to implement a scalable participatory forest monitoring system based on open-source technology that increases spatial and temporal accuracy of forest change detection, as well as the confidence in the changes detected.

1. Pratihast, A.K., Herold, M., Avitabile, V., de Bruin, S., Bartholomeus, H., Souza, C., Ribbe, L., Mobile Devices for Community-Based REDD+ Monitoring: A Case Study for Central Vietnam, Sensors 2012, 13, 21-38.
2. Pratihast, A.K., DeVries, B., Avitabile, V., de Bruin, S., Kooistra, L., Tekle, M., Herold, M., Combining Satellite Data and Community-Based Observations for Forest Monitoring, Forests 2014, 5, 2464-2489.
3. Pratihast, A.K., DeVries, B., Avitabile, V., de Bruin, S., Herold, M., Bergsmaet, A., Design and Implementation of an Interactive Web-Based Near Real-Time Forest Monitoring System, PLoS ONE 2016, 11 (3).
4. DeVries, B., Pratihast, A.K., Verbesselt, J., Kooistra, L., Herold, M., Characterizing Forest Change Using Community-Based Monitoring Data and Landsat Time Series, PLoS ONE 2016, 11(3).
5. http://geobosques.minam.gob.pe
6. Reiche, J., Verhoeven, R., Verbesselt, J., Hamunyela, E., Wielaard, N., Herold, M., Characterizing Tropical Forest Cover Loss Using Dense Sentinel-1 Data and Active Fire Alerts, Remote Sens. 2018, 10(5), 777.
7. https://landsense.eu/

The Copernicus Emergency Management Services (EMS) Mapping aims at the creation of maps and added-value information that can be used for the management of emergencies, in the early response against an event as well as in the preparation against new possible disasters and the recovery works. For these two goals, two separate modes of data creation are made available to users, the Rapid Mapping and the Risk and Recovery Mapping, respectively. Copernicus EMS Mapping Validation is used for the verification of outputs produced by those mapping services on sample basis, providing a valuable source of information to guide service improvement and evolution. This service not only aims at validating the products created by the Mapping Services, but also at evaluating the products based on the users’ feedback. The request for alternative products that could be offered by the Mapping services has led the Validation Service to the investigation of alternative methodologies or input data that could be implemented in the creation of high-quality reference data.
In this framework, the potential use of Volunteered Geographic Information (VGI) to automatically create detailed Land Use/land Cover (LULC) in an urban area that lacks of detailed LULC information and has low accessibility was assessed. Moreover, the use of social media to create reference data for validation of products was tested. For that purpose, data from the VGI platform OpenStreetMap (OSM) of a given date was analysed and compared to the LULC derived from a Very High Resolution (VHR) optical image. The applicability of social media as a source of reference data to validate LULC information was assessed using data from three different user-generated non-geographic contents, Flickr, Panoramio and YouTube. LULC data was created using an expanded version of the Urban Atlas 2006 level 4 legend. Results showed the capability of these data sources to provide information that cannot be inferred from satellite imagery. However, it should be highlighted that biases that might occur when VGI or social media are used as main input data, as an overrepresentation of sites considered interesting by the users that might not necessarily be the most relevant sites in a crisis management context might occur.

The GROW Observatory (GROW; https://growobservatory.org/), a European citizen science project, is empowering hundreds of citizens to monitor environmental variables within high-density clusters and thus to lay the foundation for scientists to validate satellite-based products at an unprecedented high spatial resolution.
Equipped with low-cost sensors and a mobile application passionate growers and enthusiastic citizens are collecting in-situ soil moisture observations to generate a reference dataset for remotely-sensed soil moisture products of unprecedented extent. Thousands of low-cost sensors are being currently deployed in nine high-density clusters distributed over Europe, establishing the basis for validating high-resolution satellite soil moisture products. To ensure reliable data collection at representative locations the growers are being trained in massive open online courses (MOOCs) and face-to-face workshops and supported through field handbooks, local community champions and an online forum. Vital for the creation of a reliable and valuable dataset is not only the quality, induced by the low-cost sensors and the citizens, but the continuity, which strongly depends on the citizens’ ongoing engagement and motivation. In contrast to conventional citizen science projects, which require only a onetime activity, GROW relies on a long time commitment by its participants to create a continuous in-situ soil moisture dataset. To this end, the GROW citizens‘ observatory framework is introduced to bridge the gap between citizens and environmental monitoring.
In this presentation we discuss the first results of validating two soil moisture products of different spatial resolution (based on ASCAT and Sentinel-1) by using a citizen-generated reference dataset. The main objective is the suitability and quality of the in-situ reference data, resulting from the low-cost sensors and the citizens participating in the observatory. First, the performance of the low-cost soil moisture sensors is determined based on calibration exercises in the laboratory and comparison with professional probes in the field. Second, the dataset properties, advantages and disadvantages as a consequence of being generated by citizens are discussed and applied methods of data preparation are described. We will demonstrate that citizens can create a dataset which is valuable for environmental monitoring and scientific applications.

In many cities the population is exposed to elevated levels of air pollution. Often the local air quality is not well known due to the sparseness of the official monitoring network, or unrealistic assumptions being made in urban air quality models (such as vehicle emission factors).

However, new sources of alternative air quality data rapidly become available. From space by satellite instruments like TROPOMI on Sentinel 5P (providing air quality information on a 7 x 3.5 km2 resolution), and on the ground by new sensor technologies allowing for low-cost and mid-cost in-situ measurements (see Figure). Numerous research groups, companies, and citizens are already experimenting with these new sensors.

The RETINA system, developed at KNMI, solves the problem how these heterogeneous data can be integrated to produce high spatio-temporal resolution maps of urban air pollution. This is not straightforward due to the localized nature of pollutants such as nitrogen dioxide (NO2). RETINA offers a two-step solution: (1) building a versatile air quality model, driven by an open source atmospheric dispersion model and emission proxies from open data sources, and (2) a practical spatial interpolation scheme, capable of assimilating observations with different accuracies.

We have implemented RETINA to map hourly NO2 concentrations on a 50m resolution in the metropolitan area of Amsterdam. Amsterdam is a mid-sized city where NO2 pollution is dominated by traffic and residential emissions. We will show that by a smart interpolation of the measurements of the official monitoring network, the air quality at undersampled areas can be estimated with reasonable accuracy. The results improve significantly when data from nearby unofficial sources (such as from low-cost sensors used in a citizen campaign [Mijling et al., 2017]) are also included.

Based on open data, the RETINA system can easily be applied to other pollutants (such as particulate matter) and to data from other sensor networks in other cities. We will show how RETINA is used to validate TROPOMI observations over urban areas. In a later stage, the satellite observations will be used as an additional data stream for assimilation, and thus provide better air quality information for cities which lack a representative official monitoring network.

Reference:
Mijling, B., Jiang, Q., de Jonge, D., and Bocconi, S.: Field calibration of electrochemical NO2 sensors in a citizen science context, Atmos. Meas. Tech., 11, 1297-1312, https://doi.org/10.5194/amt-11-1297-2018, 2018.