In the last year of the implementation of the project, our key scientific results were diffuse solar radiation prediction models based on (multilayer perceptron) neural networks. The models were constructed for various climates in Slovenia and using either only meteorological measurements, a combination of numerical weather forecasts and measurements, or only measurements. The key result that exceeded our objectives is the transferability of the models between different locations (Portorož, Maribor). Until now, solar radiation models based on artificial neural networks have only been usable for the locations for which they were trained. The transferability of the models means that we have successfully constructed models that actually capture the key physical principles. We have validated the models on longer (18 months) sets of high-quality measurement data, achieving excellent consistency evaluation parameter values (e.g. a correlation coefficient of over 0.9). The models have been placed into use and are subject to constant further development, testing and validation. This peer-reviewed article was the first paper published as part of the project. It allowed us to share our findings with our colleagues at the Slovenian IAGG association and publish the first results in the series in good time.
COBISS.SI-ID: 28378919
Numerical weather forecast in fine spatial and temporal resolution for areas above complex terrain is our key tool for predicting the spatial and temporal distribution of global solar radiation. These forecasts also serve as input data for the prediction of diffuse solar radiation using artificial neural network based models. It is therefore of vital importance to properly evaluate the performance of the numerical weather models. This article presents the specifics that need to be taken into consideration in the methodology for the validation of such models above complex terrain. It also focuses on the validation of models of wind, which is the most difficult to predict out of all the basic meteorological variables above terrain as complex as in Slovenia (apart from precipitation).
COBISS.SI-ID: 26486567
The project application also included a task in the field of information communications technology: the collection of measurement data from remote areas using DTN technology in the most difficult communication conditions (networks were set up above ground and in the Postojna Cave as part of a planned collaboration with Dr Gabrovšek of the Karst Research Institute. DTN is abbreviation for "Delay and disruption tolerant networking". Meteorological stations are usually set up in areas lacking the normal infrastructure. One of the key problems with regard to automatically captured meteorological data – which represent the input for our solar energy availability models – is also the technology for data collection from remote areas that are not covered by a communication infrastructure. Often, we are talking about remote areas where the generation of energy from solar power is so much more interesting since no other energy sources are available there or also because these are areas with a low economic value, therefore they can be utilized for e.g. extensive solar farms. In remote and inaccessible areas, the lack of communication links is a big problem especially in the site testing phase in order to verify the feasibility of the solar power utilization, when the measurements for a possibly long time period need to be obtained with minimum costs, which of course does not justify nor allow for the setup of a communication infrastructure. For communication in such harsh environments the DTN technology is being developed (in this field, MEIS has already cooperated in the EU 7OP project N4C, and in the Leonardo EU programme in its follow-up). Within this project, the data from one measurement station (Veliki Ločnik, which is a part of the testbed) is transferred by means of the DTN protocol. With the same protocol, a highly visible experiment in the karstic Postojna Cave was executed. In the published article, the researchers from MEIS and the Karst Research Institute in cooperation with other European partners have, in the example of Postojna Cave, demonstrated the usefulness of the DTN internet (a delay- and disruption-tolerant network) for the collection of automatically measured meteorological and other data in large karstic caves. According to the public opinion of reviewers, that is a “unique new method for data collection”, performed in a “highly interesting” and “appealing” way. The DTN internet is otherwise used for the most demanding of environments (from NASA’s interplanetary communication to remote areas in Lapland). The first news about the execution of the experiment in Postojna Cave also received enthusiastic praise in the form of words “THIS IS JUST GREAT!!! WHAT A STORY TO TELL!” by Dr. Vinton Cerf, who is recognized worldwide as one of the “fathers of the Internet”. Dr. Cerf is also one of the pioneers of DTN technology. The word about the article was immediately spread among researchers by Dr. Cerf himself. It certainly proves that our achievements in the DTN field have been recognized worldwide.
COBISS.SI-ID: 36767021
Our research within this project has resulted in detailed predictions of micro-meteorological characteristics of the planetary boundary layer above a complex terrain. The predictions are also highly valuable because they have been successfully validated by means of measurements. All of this has served as the foundation for the establishment of a broader collaboration between MEIS and the Department for Public Health of the Ljubljana Faculty of Medicine as well as the Ljubljana Institute of Oncology. The first topic deals with respiratory diseases (in addition to this joint article with the aforementioned two teams, we have presented another paper at a conference in Slovenia, while a third article has been also published in an international journal). We are also already developing methodologies for studying the spatial dependency of the impact of solar radiation (for which we have collected several years of data by components as part of this project) on skin cancer in collaboration with the Institute of Oncology, which constitutes a continuation of the method outlined in this article. This will also most likely be the key continuation of the project.
COBISS.SI-ID: 1824635
The article focuses on all the aspects of weather and air pollution prediction at a regional level using the QualeAria system. Forecasts of global solar radiation and air pollution with PM10 (particles that cause the dimming and scattering of direct solar radiation) serve as the key input data for solar radiation prediction at a local level in a fine spatial and temporal resolution. Solar radiation is important for energy generation and is also a natural driving force that determines the air circulation processes – especially vertically – during the day, which is important for the reduction of air pollution. The article provides a detailed evaluation of the performance of the prediction system. Based on these findings, we have also developed special display tools for evaluating the correlation between solar radiation and the concentrations of pollutants in ambient air (measured or modelled).
COBISS.SI-ID: 28053799