Development of Global NWP Model: KIAPS-GM

Emilia Kyung Jin

Korea Institute of Atmospheric Prediction Systems (KIAPS)

Korea Meteorological Administration (KMA) launched a 9-year project in 2011 to develop Korea's own global NWP model system. To lead the effort, Korea Institute of Atmospheric Prediction Systems (KIAPS) was established by KMA as a non-profit foundation. In 2013, the initial version of Korea Institute of Atmospheric Prediction Systems-Global Model (KIAPS-GM) is developed. It uses the 3-D hydrostatic dynamical core of HOMME (High Order Method Modeling Environment) based on combined spectral element, which is highly scalable, of high order. So it has an excellent scalability under parallel computing environment, and has no pole singularity problem. It uses the finite different vertical discretization, hybrid sigma vertical coordinate, and Lorenz grid for physics. The physical package of KIAPS-GM consists of schemes of radiation, surface layer, land surface, boundary layer, cumulus convection, large-scale cloud, cloud microphysics, orographic and nonorographic gravity wave drag process. A scheme of each physical process has been selected based on performances, technology trends, and coupling facilities of existing schemes based on the comprehensive analysis of world's leading operational models and state of the art climate models. Both dynamical core and physics package is coupled via KIAPS's own modeling framework. The initial version of KIAPS-GM has the horizontal resolution of about 100km by 100km in cubed sphere model grid (ne30np4 resolution) and the model top of 0.003 hPa with 70 layers. A set of ensemble experiments were conducted to validate model.

Current status of land surface model in KIAPS-GM

Kyung Hee Seol and Emilia Kyung Jin

Korea Institute of Atmospheric Prediction Systems (KIAPS)

The land surface model (LSM) in numerical weather prediction or climate general circulation model is important mainly because the sensible and latent heat fluxes at the surface are the lower boundary conditions for the energy and moisture equations in the atmosphere. For the KIAPS-GM LSM, the Noah LSM, developed by NCEP, OSU, Air Force, and Hydrology Research Lab. at the NWS, was chosen. The source code was extracted from the GRIMs and implemented in the KIAPS-GM framework following the coupling strategy. For the vegetation and soil ancillaries, the SiB dataset used in the NCEP GFS and GRIMs are used. Because the KIAPS-GM is developed based on the cubed sphere, the ancillaries on lat-lon grid are converted into the cubed sphere grid using the SCRIP conservative remapping method.

The off-line and on-line tests with KIAPS-GM were conducted for 1-day and 15-day period, started from 00Z 1 January 2012, respectively. The dynamic and surface initial data are from the ERA-interim and sea surface temperature and sea-ice climatology from the HadISST are used. Compared to the reanalysis, the overall distributions of the simulated surface fields seem to be reasonable for both off-line and on-line tests. The longer simulation with KIAPS-GM will be conducted for the detailed evaluation.

Impacts of soil moisture initialization on operational global NWP model forecasts

Kyoungmi Cho¹, Kyung-Hee Seol¹, Emilia Kyung Jin¹, Paul Dirmeyer², Zichang Guo²

¹Korea Institute of Atmospheric Prediction Systems (KIAPS)

²Center for Ocean-Land-Atmosphere Studies (COLA)

Soil moisture is a key physical variable in controlling the exchange of water, energy fluxes between the land surface and atmosphere in regional and global scales. It thus affects near-surface climate by changing soil property. Soil moisture plays a some important role in the formation and development of meso- and micro-scale weather system. The spatial variation of soil moisture may affect baroclinic structure of the lower tropospheric atmosphere.

To initialize the land surface model, KIAPS is developing the retrospective land surface analysis technique for the soil moisture in collaboration with COLA group. In 2013, as a test, two sets of numerical weather prediction experiments were conducted with different land surface initial states for Unified Model (UM). Soil moisture obtained from Global Land Data Assimilation System (GLDAS) was processed for simulating UM, and results are compared with UM control run with its original surface process (SURF) and observation as well.

Initial soil moisture effects both surface temperature and precipitation forecast in this experiment and GLDAS test show improved result to reduce RMSE and bias for each variable. However, because the impact varies according to region, we will carry out more detailed analysis for each area.

Calibration, compensating errors and data-based realism in land surface models

Gab Abramowitz

Climate Change Research Centre

University of New South Wales

In an environment where model skill has well defined metrics and spatial scales of application, the issue of compensating errors within the modelling system may be of little consequence. That is, the ability to calibrate a model against known metrics, thus positioning a model further toward being empirically based rather than physically-based, can be an advantage in terms of performance gains. The climate modelling community, however, is beginning to come to terms with a transition from an environment where models are created as hypothesis testing tools with specific variables and metrics in mind to one where the totality of model simulation is fair game for evaluation. This transition represents a tremendous leap in scope and responsibility. This talk will highlight a few aspects of land surface model application that present potential issues for this transition and argue that if the benefits of seamless prediction are to be realised, the competing operational pressures of forecasting and climate projection need to be reconciled. In particular, an emphasis on data-based realism and avoidance of compensating errors is likely key. Examples of questions that highlight this issue include: Are land surface models over-parameterised when fewer than 5 key parameters are identifiable at the usual scale of their application? How would two apparently equally well performing models with low versus high coupling strengths perform if implemented at a range of spatial scales? Does the acceptance of soil moisture as a model-specific quantity affect skill related to land surface memory when the time and space scales of model application change? This presentation will not answer any of these questions, but aim to set a context for discussion of more specific issues within the three workshop themes.

Scaling the heterogeneously heated convective boundary layer

Chiel van Heerwarden

Max Planck Institute for Meteorology

Scaling laws for the heterogeneously heated free convective boundary layer (CBL) are presented based on dimensional analysis and results from large-eddy simulations (LES) and direct numerical simulations (DNS). A series of simulations has been done in which a CBL forms in a linearly stratified atmosphere that is heated from below by square patches with a high surface buoyancy flux. Each simulation has been run long enough to contain first the formation of a peak in kinetic energy, corresponding to the "optimal" heterogeneity size with the strongest secondary circulations, and second the subsequent transition into a horizontally homogeneous CBL.

The results show that the optimal state and the transition do not occur at a fixed ratio of the heterogeneity size and the CBL height, but instead take place relatively earlier for larger heterogeneity sizes, caused by the organization of the downward moving velocity that breaks down the secondary circulations. The time instance of the occurrence of the optimal state and transition are related to the heterogeneity amplitude: stronger amplitudes have an earlier optimal state and a later transition. Our results suggest that larger scale heterogeneity is able to enhance entrainment.

In order to introduce the effects of heterogeneity into large scale models it is of equal importance to have information on the heterogeneity size and of the heterogeneity amplitude.

Emergent Pathways of Predictability - The Diurnal Cycle and the Role of Landform and Landcover on the Organization of Moist Processes at Multiple Scales

Ana Barros

Duke University

Results from the analysis of remote-sensing and ground-based observations, and model simulations point to the ubiquitous persistence of space-time co-organization of landform, precipitable water, aerosols, clouds and precipitation systems amidst nonlocal multi-scale climate variability. Here, the focus is on the diurnal cycle as a necessary condition of sub-seasonal predictability. Using nonlinear metrics and scaling analysis, the question of hydrometeorological predictability is addressed in three ways: 1) high resolution numerical simulations, 2) data-driven place-based long-range operational forecasting and downscaling of model simulations; and 3) observations. First, we demonstrate how landform and landcover conspire to organize spatially the diurnal cycle of clouds and precipitation in mountainous regions. Second, we contrast model simulations and observations and use various diagnostics to identify key boundary-layer physics, local versus remote processes, and model artifacts (parameterizations and numerics) in determining the space-time dynamics of the diurnal cycle over different types of landscapes. Recent research pointing toward a physical-statistical scaling framework to improve the parameterization of moist processes in atmospheric models across scales will be presented.

Land Surface model Scaling issues over West Africa: Perspectives from the AMMA Land surface Model Inter-comparison Project (ALMIP)

Aaron Boone

Mètèo-France

The AMMA (African Monsoon Multidisciplinary Analysis) project was organized in recent years with the main goal of obtaining a better understanding of the intra-seasonal and inter-annual variability of the west-African monsoon (WAM). Land surface processes have been shown to be strongly coupled to the atmosphere on various temporal and spatial scales over West Africa, therefore a high priority goal of AMMA was to better understand and model the influence of the spatio-temporal variability of surface processes on the atmospheric circulation patterns and the regional water cycle. This is being addressed through a multi-scale modelling approach using an ensemble of land surface models (LSMs) which rely on dedicated satellite-based forcing and land surface parameter products, and data from the AMMA observational field campaigns. The idea is to force state-of-the-art land surface models with the best quality and highest (space and time) resolution data available in order to better understand the key processes and their corresponding scales within the context of an Land Surface Model (LSM) inter-comparison project. The ultimate goal is to improve the representation (including the effects of scaling) of key surface, vegetation and hydrological processes, which can then in turn benefit the forecast and regional and global climate modelling communities.

In the now completed ALMIP phase 1, LSMs were run over all of West Africa for a several year period using a typical RCM grid resolution (0.5 deg) in order to obtain regional scale estimates of key surface processes and state variables. In contrast, ALMIP phase 2 deals with the local to meso scales (0.05 deg). LSMs are being forced and evaluated using observational data from three heavily instrumented supersites from the AMMA-Couplage de l'Atmosphère Tropicale et du Cycle Hydrologique (CATCH) observing system which covers a north-south transect encompassing a large eco-climatic gradient. In this talk, the impact of scaling on atmospheric variables and land surface/hydrological processes will be highlighted. The impact of precipitation forcing which more closely resolves the convective nature of the precipitation (compared to usual forcings based on 3-hourly data) will be presented. In addition, at the local and meso scales, certain key processes come to light which must be modelled which were not evident at larger scales. These are related to the endoric hydrological processes in the semi-arid context. It was also found that deep rooting vegetation is able to tap near-surface slowly-varying water tables, which implies a much longer land surface memory effect than is currently represented in coupled models.

Multiscale land surface impacts on regional weather and climate

Dev Niyogi

Purdue University

Land use/cover change is the other global change underway. Land surface modulates portioning of surface energy which in turn affects the surface boundary layer response, regional mesoscale convergence and the associated mesoscale convective triggers and circulation patterns. These feedbacks with sufficient spatial intensity can provide organized atmospheric feedbacks and affect regional climatic patterns. Yet detecting these patterns and feedbacks from observations has been difficult due to various interactive processes actively involved. This presentation will review the state of the science in identifying these feedbacks using results from recent observational and model synthesis studies. Results will focus on discussion and diagnosis of landscape change and land heterogeneity feedbacks on regional weather and climate and provide suggested pathways for improvements in our future projections and assessments of high impact weather and regional climate change.

Land-surface controls over evaporation variance

Randal Koster

NASA Goddard Space Flight Center

Soil moisture variability influences atmospheric variability through its impact on the surface energy and water budgets – through its control over the partitioning of precipitation water into evaporation, runoff, and water storage and the partitioning of incoming radiative energy into latent heat flux, sensible heat flux, and heat storage. Given the connection between evaporation and latent heat flux, evaporation variability lies at the very heart of identifiable land-atmosphere feedback. Here, using a simple water balance model, we examine the sensitivity of evaporation variance to different combinations of land model evaporation and runoff formulations, identifying the combination that produces results most consistent with available observations.

Representing natural land surface heterogeneity in Numerical Weather Prediction and Earth System models: benefits and limitations of current schemes in presence of large contrasts

Gianpaolo Balsamo

European Centre for Medium-range Weather Forecasts

Abstract: Natural and anthropogenic variability that characterises the Earth's surface is very much a driver for complexity and increased resolution in NWP and ESS models, with existing or upcoming support from Earth-Observation datasets such as those derived from MODIS, SPOT, PROBA-V, Sentinels ESA's program satellites, all attaining sub-kilometer remote-sensing capabilities.

While a higher resolution improves the description of contrasting surfaces such as land-water or snow-forest, large part of these contrasts is bound to remain a sub-grid parameterisation's issue due to fractal nature of the surface: coastlines, rivers, lakes, forests, urban-areas are rarely pure land-use categories even when approaching kilometric scales. Representing those contrasts is important for partitioning energy and water fluxes but also for carbon-exchange and therefore have received attention both in NWP and ESM. The land surface tiling is often a method used for such purposes as it can accommodate several parameterisations for the different surface types. We have examined the capacity of the tiling to represent large natural contrasts in presence of snow, forests and lakes, in the framework of the land surface scheme operationally used at ECMWF. The benefits on representing the fluxes and the limitations coming from lateral decoupling (lack of mixing) are highlighted in a set of field-site examples and the impact is evaluated in global sensitivity experiments. Preliminary ideas to go beyond the tiling concepts will be discussed.

Impact of soil moisture initialization on AMIP type simulations in JSBACH

Tobias Stacke & Stefan Hagemann

Max Planck Institute for Meteorology

Recently, the land surface model JSBACH was improved by replacing its bucket-type soil hydrology scheme with a multi-layer scheme. This new scheme is a more realistic representation of the soil including percolation and diffusion fluxes between up to five separate layers, the limitation of bare soil evaporation to the uppermost soil layer and the addition of a long-term water storage below the root zone in regions with deep soil. While the hydrological cycle is not strongly affected by this new scheme, it has some implications on the simulated soil moisture memory, which is mostly strengthened due to the additional deep layer water storage.

Furthermore, this scheme allows for the assimilation of soil moisture satellite observations. Such observations are limited to the uppermost part of the soil and are best represented by a layered soil moisture scheme, which included such a thin upper soil layer. In order to evaluate whether the assimilation of soil moisture might have a significant impact on the simulated water and energy cycles, we conducted a number of AMIP-type simulations where the initial soil moisture state is changed towards representative very dry or very wet conditions. We investigate the impact of the initialization on global scale and try to derive conclusions about the persistence of soil moisture perturbations for different seasons. These will be compared to soil moisture persistence analyses based on the autocorrelation of soil moisture time series.

Noah-MP: A New Paradigm for Land Surface Modeling

Zong-Liang Yang¹, Fei Chen², Mike Barlage², Mike Ek³, Guo-Yue Niu⁴

¹University of Texas at Austin; ²National Center for Atmospheric Research; ³National Centers for Environmental Prediction; ⁴University of Arizona

Multi-physics (or multi-parameterization) frameworks have been incorporated in land surface models to allow for multi-hypothesis testing, uncertainty quantification, and ensemble climate and hydrological predictions at intraseasonal to decadal timescales. One such example is the community Noah land surface model with multi-parameterization options (Noah-MP), which augments the standard Noah land surface model, capable of producing thousands of combinations of parameterizations, in addition to its improved realism and enhanced capabilities for seasonal prediction (multi-layer snowpack, groundwater dynamics, and vegetation dynamics). Noah-MP has been linked to the Weather, Research and Forecasting model and the National Centers for Environmental Prediction Climate Forecast System. This talk will briefly review the Noah-MP framework, highlight its recent results from offline and coupled WRF/Noah-MP simulations, and provide a prospect for new applications.

The impact of land surface on sub-seasonal atmospheric prediction and its variability

Zhichang Guo

Center for Ocean-Land-Atmosphere Studies

Contribution of land surface initialization to sub-seasonal atmospheric predictability and prediction skills not only varies with season, but also has interannual variability. This study examines the temporal variability of land surface impacts on atmospheric prediction and its association with variations of land-atmosphere coupling strength and soil moisture memory time. Also, evidences for the impacts are explored from observations of precipitation, evaporation, soil moisture, and near-surface air temperature.

Coupling of Diurnal Climate to Clouds, Land-use and Snow

Alan Betts

Atmospheric Research

Hourly observations from 1953-2011 of temperature, RH, opaque cloud cover; and precipitation and snow depth from 14 climate stations across the Canadian Prairies together with ecodistrict crop and MODIS data to analyze the coupling of the diurnal climate to clouds, land-use and snow.

1) The cloud forcing of the diurnal climate has distinct warm and cold season behavior. From April to October, when incoming shortwave radiation dominates over longwave cooling, maximum temperature and the diurnal ranges of temperature and relative humidity increase with decreasing opaque cloud cover, while minimum temperature is almost independent of cloud. During the winter period, both maximum and minimum temperature fall with decreasing cloud, as longwave cooling dominates over the reduced net shortwave flux with snow.

2) The agricultural land-use conversion from summerfallow to annual cropping on 5 MHa (15-20% of the land area in Saskatchewan) in recent decades has cooled and moistened the summer climate due to increased transpiration, increasing cloud cover and precipitation.

3) The fall-winter and winter-spring transitions in November and March between warm and cold seasons occur within days of snowfall and snowmelt, with a change of 2-m temperature of 10K. 10% fewer days with snow cover gives a winter climate warming of 1.2 to 1.5K.

Land-atmosphere interaction and cloud formation

Michael Ek

NOAA National Centers for Environmental Research / Environmental Modeling Center

The interaction between the land and atmosphere affects the evolution of both these systems, and requires a proper understanding and then realistic representation of the physical processes in weather and climate models. This interaction begins at a local scale, i.e. between surface fluxes and the near-surface atmosphere, with the strength of near-surface coupling dependent on the relationship between soil moisture and evapotranspiration. This relationship is a function of the surface-layer turbulence and vegetation and soil properties, where for strong (weak) coupling, a given change in soil moisture yields a large (small) change in evapotranspiration. Subsequently, land-atmosphere interaction involves the evolution of land surface fluxes and atmospheric boundary (ABL) development as a coupled system, where a number of competing process e.g. lead to the formation of ABL clouds (i.e. fair weather cumulus).

Modelled contrast in the response of the surface energy balance to heatwaves for forest and grassland

Bart van den Hurk

Royal Netherlands Meteorological Institute (KNMI)

Observations have shown that differences in surface energy fluxes over grasslands and forests are amplified during heat waves. The role of land atmosphere feedbacks in this process in still uncertain. In this study, we use a single-column model (SCM) to investigate the difference between forest and grassland in their energy response to heat waves. Three simulations for the period 2005-2011 were carried out: a control run using vegetation characteristics for Cabauw (the Netherlands), a run where the vegetation is changed to 100% forest, and a run with 100% short grass as vegetation. A surface evaporation tendency equation is used to analyse the impact of the land atmosphere feedbacks on evapotranspiration and sensible heat release under normal summer and heatwave conditions with excessive shortwave radiation.

Land atmosphere feedbacks modify the contrast in surface energy fluxes between forest and grass, particularly during heat wave conditions. The surface resistance feedback has the largest positive impact, while boundary layer feedbacks generally tend to reduce the contrast. Overall forests give higher air temperatures and drier atmospheres during heat waves. In offline land surface model simulations the difference between forest and grassland during heat waves cannot be diagnosed adequately owing to the absence of boundary layer feedbacks.

Stable boundary Layers and Diurnal Cycles - Challenges for Weather and Climate Models

Bert Holtslag

Wageningen University

The representation of the atmospheric boundary layer is an important part of weather and climate models and impacts many applications such as air quality and wind energy. Over the years, the performance in modeling 2 m temperature and 10 m wind speed has improved but errors are still significant. This is in particular the case under clear skies and low wind-speed conditions at night as well as during winter in stably stratified conditions over land and ice. In this paper, we review these issues and provide an overview of the current understanding and model performance. Results from weather forecast and climate models are used to illustrate the state of the art, as well as findings and recommendations from three inter-comparison studies held within the "Global Energy and Water Exchanges (GEWEX)" Atmospheric Boundary Layer Study (GABLS). Within GABLS, the focus has been on the examination of the representation of the stable boundary layer and the diurnal cycle over land in clear sky conditions. For this purpose, single-column versions of weather and climate models have been compared with observations, research models and Large Eddy Simulations. The intercomparison cases are based on observations taken in the Arctic, Kansas and at Cabauw in the Netherlands. From these studies, we find that even for the non-cloudy boundary layer important parameterization challenges remain.

Land-atmosphere interactions: from local to synoptic scale

Pierre Gentine

Columbia University

In this presentation we will show how nonlinearities in the land-atmosphere coupled system arise, with a focus on the boundary layer, clouds, precipitation and feedback to synoptic circulation. We will first show that because of non-linear irises induced by Clausius Clapeyron both negative and positive soil moisture-cloud/precipitation feedback can be observed locally. These feedbacks provide valuable diagnostics of the surface evaporative fraction and remote sensing of cloud occurrence can be used to estimate the surface energy budget. Finally these local feedbacks are coupled through the larger scale circulation. We will finally show how the complementary relationship and the Budyko curve naturally arise from those synoptic scale interactions.

When is the Land Surface Important for Triggering Convection?

Ahmed Tawfik

Center for Ocean-Land-Atmosphere Studies

The atmospheric component of the earth system is described by rapid hourly and daily fluctuations in typical state variables, whereas terrestrial state variables evolve on daily to yearly timescales. Here the ability of the land surface to trigger convection is explored from diurnal to seasonal timescales using a new process-based framework, the Heated Condensation Framework (HCF). This framework allows for the quantification of how preconditioned the atmosphere is to convection, and is captured by a single conserved quantity called the buoyant condensation level (BCL). Using the BCL, the impacts of the surface fluxes on convective triggering potential can be separated from other factors controlling convection, such as advection and entrainment. The scale of these interactions will be explored in observations and models.