Two types of applied tasks used in the thematic virtual research environment (VRE) based on the “Climate+” platform are considered. Tasks of both types use significant amount of climatic or meteorological data. The first type of applied tasks whose solutions describe quantitatively climate of chosen territory are on-line computed and mapped using GIS technologies. The second type of applied tasks includes tasks used for decision making. Those along with the computational component, includes tools for expert selection of the initial conditions for these tasks, tools for determining the semantic homogeneity of physical quantities used in the calculations, and software for forming the A-box of the knowledge base of a decision support system (DSS). Presented are several first-type tasks and the second-type task about changing the depth of the active soil layer in the northern regions of Western and Eastern Siberia (the interfluve of the Ob and Yenisei rivers) for a period of 60 years. The solution of this task and the structure of a typical ontology individual used for the decision making are presented. The role of the ontology description of solutions of applied tasks in the VRE based on the “Climate+” platform is discussed.
The creation of a virtual research environment (VRE) to
deal with large data arrays becomes popular in data
intensive domains [
Necessity of ontologies usage in geophysical sciences and their role have been demonstrated in the papers [59].
In this paper we describe new steps in development of the ‘Climate+’ platform, namely applications of different types and ontology that refer to three layers (Data, Metadata (Information) and Ontology (Knowledge) layers) of data processing services of the VRE under development. Then we describe the Ontology layer and the applied problem. Special attention will be paid to semantic heterogeneity and formalization of thematic domains related with applied tasks, in particular, a solution of a reduction problem.
There are two types of tasks which should be solved
using the thematic VRE. The first type tasks are simple
tasks of calculation of some physical value
characterizing climatic processes. Examples of those
solved within the previously developed “Climate”
platform can be found in paper [
The report discusses the ways of forming the DSS knowledge base using OWL-ontologies that describe solutions of applied tasks and their properties.
The web GIS platform ‘Climate+’ developed at the
IMCES SB RAS is aimed at processing and analysis of
geospatial gridded datasets in Earth system science, and
online visualization of results [
The client part of the architecture is based on modern graphical web browser. It is presented by a single ‘Client applications’ tier, respectively.
The data layer contains netCDF datasets and PostGIS databases while the metadata layer represents the Metadata database (MDDB) describing geospatial datasets and their processing routines framework of the computational backend.
The computational backend developed contains data processing and visualization software components based on GNU Data Language (GDL, http://gnudatalanguage.sourceforge.net/) and Python. Geospatial datasets are processed by a specialized set of validated software modules running within the The visualization component of the backend generates files in the following formats: GeoTIFF, ESRI Shapefile, Encapsulated PostScript, CSV, XML, netCDF, float GeoTIFF. The final results are represented by raster and vector cartographical layers accompanied by corresponding binary netCDF data.
The geoportal provides cartographical web services such as WMS, WFS, WPS, as well as server-side part of the Web-GIS client applications which comply with general INSPIRE (INfrastructure for SPatial InfoRmation in Europe, https://inspire.ec.europa.eu) requirements to geospatial data visualization.
The results of the first type applied tasks (computational problems describing changes of states of spatio-temporal objects) solving are added to the data and metadata layers and might be used later. To solve the problem of semantic heterogeneity in the ‘Climate+’ platform, an ontology layer characterizing the properties of the data
collections (Reanalysis, Observations, Modelling Data) is created. This ontology is used to select input data for Applied Tasks applications.
To describe extreme climate events statistics dedicated
analytical software tools were integrated into the web
[
At present, the system allows to calculate basic statistical characteristics and indicators of the temporal structure of meteorological series, describing patterns of changes in time and space. The functionality of the system includes a calculation of trends, an assessment of their statistical significance and a degree of correlation of meteorological quantities. The IPCC recommended climatic change indices are also calculated: extreme values of daily temperature and daily rainfall and their probabilistic characteristics. In addition, the system calculates characteristics of nonstationary distributions of various extreme values. The implemented set of computational procedures makes it possible to get a complete picture of peculiarities of occurring changes in climatic characteristics for the region of study. Modular organization of the
system functionality by adding new allows to software expand its components developed by both developers and users of the system.
A statistical description of extreme precipitation and temperature can be achieved using the concepts of extreme value statistics (EVS).
Software implementation of EVS in
R language
(package "extRemes") allows statistical modelling of
maximum values based on a non-stationary generalized
extreme value distribution. Required for risk assessment
quantities can be calculated using this distribution
function. In particular, it is the probability of the
observed variable to exceed a certain level. These levels
are frequently expressed as return levels for a certain
return period . is defined as the level which is
exceeded on average every , i.e., with probability .
This functionality was used to calculate 100-years return
levels of July maximum precipitation based on ECMWF
ERA Interim [
Figure 2 100-years return levels of July maximum precipitation for the Southern Siberia region: a ECMWF ERA Interim data, 0.75x0.75 horizontal grid, 1979-2007 years, b - APHRODITE JMA data, 0.25x0.25 horizontal grid, 1979-2007 years Results obtained demonstrate similar behavior of the calculated characteristic, but the APHRODITE JMA results have more details and higher values in some regions.
The analysis of trends in meteorological observations is one of the most common activities in climate change studies. Quantile regression provides a well-defined statistical framework for estimating the rate of change not only in the mean as in ordinary regression, but in all parts of the data distribution. Given a random variable with cumulative continuous distribution function ( ), the quantile function ( ) is defined from the ( ) as ( ) = −1( ). The quantile is defined as the value ( ) such that [
≤ ( )] = , 0 ≤ ≤ 1. Then, considering the conditional distribution of given = , the conditional quantile function | ( ; ) verifies [ ≤ | ( ; )| = ] = .
Whereas ordinary regression is based on the conditional mean function [ | = ] and minimization of the respective residuals, quantile regression is based on the conditional quantile function and minimization of the sum of asymmetrically ∑ =1 ( )| − | ( ; absolute value function.
weighted absolute
residuals
= )|, where is the tilted
Quantile regression calculation is implemented in R
language by the software package "quantreg" [
The computational backend developed contains data processing and visualization software components based on GNU Data Language (GDL, http://gnudatalanguage.sourceforge.net/) and Python. Geospatial datasets are processed by a specialized set of validated software modules running within the framework of the computational backend.
Based on ECMWF ERA 40 data [
Results obtained show that maximum January temperature at quantile 0.05 is changed (both decreased and increased) to the greater extent in comparison with (b) and (c) almost everywhere. Temperature at quantile 0.95 is changed to the less extent.
The general formulation of the task is considered in a
number of publications [
T z H : 0, t 0.
z
Here the functions φ(z) and χ(t) are the calculation results
in the soil model of Institute of Numerical Mathematics
RAS. Our task statement is a simpler than the statement
suggested earlier in the work [
Output data of the model are monthly mean vertical profiles of the soil temperature from the surface to a depth of 20 m on an inhomogeneous vertical grid (0.01, 0.02, 0.04, 0.08, 0.15, 0.25, 0.35, 0.45, 0.55, 0.65, 0.75, 0.85, 0.95, 1.05, 1.15, 1.25, 1.35, 1.45, 1.55, 2.0, 3.0, 5.0, 10.0, and 20.0 m) in each cell of a horizontal grid. Several Circumpolar Active Layer Monitoring (CALM, [https://www2.gwu.edu/~calm/data/north.html]) stations which measure soil temperature and frozen depth are situated in Western Siberia (Table 1). This resource provides for annual data on the depth of frost penetration in the soil (see) by the end of the thaw season.
PGP, for Parisento, Gydan Peninsula; VDYP, for Vaskiny Dachy, Yamal Peninsula; UGF, for Urengoy Gas Field; and HPU, for Harp, Polar Urals.
The results of the numerical solution of the task are numerical arrays of temperatures at different depths (24 levels) for 720 months (60 years) for each of 100 horizontal cells (10 in latitude and 10 in longitude). Based on the values of the numerical arrays, parameters p1 and p2 are calculated, which are the values of the properties of the applied ontology individuals generated. The properties of these individuals are listed in Table 2. Based on the numeric arrays, the characteristics selected, which are the values of the properties of applied ontology individuals generated, are calculated. The property t3:hasAnnual_End-of-Season_Thaw_Depth is valuable only in the case where the measurement station coordinates fall in a 1.5°×2° cell centered at a point that corresponds to values of the properties t3:hasLatitude and t3:hasLongtitude. No individuals are generated for cells with negative soil temperatures.
Two types of individuals are constructed by 1728000 soil temperature values calculated. Examples of the structure of such individuals are given in Figs. 5 and 6. Using these temperature values, its annual means are calculated, and the number of a level at which the soil temperature changes sign is found. The value of the maximum thaw depth allows one to compare the numerical simulation results with the measurement data. The mean temperature and depth values can be used to analyze the permafrost structure (transition from continuous to discontinuous propagation) and the character of the phenomenon (periodicity or trend). FTS is the prefix that shows that an individual of the solution of Freeze-up and Thawing of Soil task; <x,y> are the geographical coordinates of the center of a computing cell; <year> is the year; <month> is the month; the ending Depth/Soil_Temperature designates which physical parameter is described.
We presented description of the two types of application used 80 Tb IMCES SB RAS collections of climatic and meteorological data. In the first type of applications statistical characteristics of climatic characteristics are on-line calculated and relevant results are presented to user as maps of their fields. The second type applications are aimed at short- or long-term prognosis of physical values evolution important for decision making. As an example of such application the problem of long-term active soil layer evolution in Northern regions is considered. On the basis we created the database which is used to form knowledge base about evolution of physical values determining behavior natural objects and a planned transport infrastructure in the Northern part of West Siberia.
To expand web GIS platform ‘Climate+’ functionality we plan development of a new software module utilizing the powerful package "copula". It will allow one to calculate probability distributions of multivariate random variables and determine a structure of dependence between different climatic variables. Acknowledgments. The authors thank the Russian Science Foundation for the support of this work under the grant No16-19-10257.