5th iLEAPS Science Conference Abstracts - C4

Abstracts – Session C4

Dryland ecosystems: New modelling and measurement challenges

C401 -Invited: Dry-land ecosystem processes related to water, carbon and energy under anthropogenic pressure

Dan Yakir

C402ORAL-0416: The Complexity of the interaction between drought, rainfall timing and rising atmospheric CO2

Sebastian Leuzinger1, Mark Hovenden1, Simone Fatichi2

1Institute for Applied Ecology New Zealand, School of Science Auckland University of Technology, Auckland, New Zealand 2Institute of Applied Ecology New Zealand, School of Science Auckland University of Technology, Auckland, New Zealand

In this talk I am going to address and challenge two generally accepted patterns from the past decades of global change experiments and modelling: (1) Rising atmospheric CO2 stimulates biomass production primarily via stimulation of photosynthesis, and, to a lesser extent, via an indirect effect arising from CO2-induced water savings. (2) The indirect effect of biomass stimulation is generally higher at drier sites. Using results from a mechanistic modelling study, I will first show that indirect CO2 effects are indeed more important in dry ecosystems and can outgrow direct effects on leaf photosynthesis. The exact mechanisms of these indirect effects can only be explored using models, as direct and indirect CO2 effects will always be confounded in experiments. Conversely, using a meta analysis across eleven globally distributed CO2 experiments in grassland, it appears that while dryness is important for the CO2 response, the latter is heavily influenced by seasonal rainfall timing: sites with higher spring precipitation showed a consistently higher CO2 response. This suggest a delicate balance of dominating direct vs. indirect CO2 fertilisation effects, driven by seasonal rainfall timing. At a finer temporal resolution, such switches from direct to indirect CO2 effects remain speculative, but may be much more complex than implemented in our current models.

C403 - ORAL-0033: Plant functional diversity affects simulated climate-vegetation interaction in semi-arid regions

Vivienne Groner1, Thomas Raddatz1, Christian Reick1, Martin Claussen1, 2

1Max Planck Institute for Meteorology, Hamburg, Germany 2Center for Earth system research and Sustainability, University of Hamburg, Hamburg, Germany

We present a series of coupled land-atmosphere simulations with different combinations of plant functional types (PFTs) from mid-Holocene to preindustrial to show how plant functional diversity affects simulated climate-vegetation interaction in subtropical Africa.
 
Scientists nowadays agree that the establishment of the ``green'' Sahara was triggered by external changes in the Earth's orbit and amplified by internal feedback mechanisms. The timing and abruptness of the transition to the ``desert'' state are in turn still under debate. While some studies indicated an abrupt collapse of vegetation implying a strong climate-vegetation feedback, others suggested a gradual vegetation decline thereby questioning the existence of a strong climate-vegetation feedback. However, none of these studies explicitly accounted for the role of plant diversity.
 
We show that the introduction or removal of a single PFT can bring about significant impacts on the simulated climate-vegetation system stability and the system response to changing external forcing. While simulations with the standard set of PFTs show a gradual decrease of precipitation and vegetation cover over time, the reduction of plant functional diversity can cause either an abrupt decline of both variables or an even slower response to the external forcing. PFT composition seems to be more important for the system response to external forcing than the total number of PFTs, and an increase in plant functional diversity does not necessarily increase the stability of a climate-vegetation system.
 
From this we conclude that accounting for plant functional diversity in future studies - not only on palaeo climates - could significantly improve the understanding of climate-vegetation interaction in semi-arid regions, the predictability of the vegetation response to changing climate, and respectively, of the resulting feedback on precipitation.

C404ORAL-0292: Can the Status of Grasslands in Semi-Arid Regions be Defined by Regression Parameters?

Catherine Van den Hoof1

1Global Change Institute, University of the Witwatersrand, Johannesburg, South Africa

Several studies have highlighted the lack of an appropriate quantitative definition of vegetation status, in particular for grassland ecosystems, in the context of land degradation. In semi-arid regions a large fraction of the variability in grassland productivity can be explained by precipitation variability, due to the significant role of water availability on productivity. Both productivity and precipitation are assumed to be positively linearly correlated between 400 and 800 mm of precipitation per year, however with a stronger correlation across space than through time due to a lag in response of ecosystems to changes in water availability.
 
The grassland ecosystem status could therefore be described by the mean productivity given the mean precipitation and by the sensitivity of the grassland to the inter-annual variation in precipitation. The predicted productivity and sensitivity could further fluctuate within an interval given site specific characteristics such as soil type. In this study, this hypothesis is evaluated at 60 grassland sites within South Africa based on the monthly FAPAR from the high resolution MISR-HR satellite data and the CHIRPS monthly precipitation dataset. The grassland sites have been selected in such a way that they are well representing the 400-800 mm precipitation range. The overall objective of this research is to define quantitatively the status of grassland in semi-arid regions, as a norm to be used in the context of land degradation, based on the relationship of grassland productivity with precipitation mean and variability.

C405 - ORAL-0275: Soil moisture memory and early warning of agricultural drought in semi-arid Africa

Emily Black 1, Matt Brown2, Dagmawi Asfaw1, Tristan Quaife1, Ewan Pinnington1, Fred Otu-Larbi3

1University of Reading, Reading, United Kingdom 2University of Oxford, Oxford, United Kingdom 3Ghana Meteorological Agency, Accra, Ghana

Farmers in Africa are highly vulnerable to variability in the weather – especially to drought.  The risk of agricultural drought is affected by the properties of the land surface, the antecedent soil moisture and the ris of rainfall deficits.  Robust and timely information on drought risk can enable farmers to take action to increase yields. We present a new agricultural decision support framework, TAMSAT-ALERT (TAMSAT-AgricuLtural EaRly warning sysTem), which uses a process-based land-surface model, JULES, to derive agricultural risk assessments from meteorological observations and forecasts, combined with information about the historical climatology and the properties of the land surface. In addition to its operational utility, assessment of TAMSAT-ALERT hindcasts provides new insight into variability and change in soil moisture memory.
 
This presentation will describe how better understanding of soil moisture memory in semi-arid regions can improve early warning and hence build the resilience of farmers in Ghana to drought.

C406 POSTER-0064: Monitoring dryland energy and water dynamics in India: an analysis of COSMOS-India and flux tower observations

Hollie Cooper1, Jenna Thornton1, Ross Morrison1, Jonathan Evans1, Christopher Taylor1, Lucy Ball1, Alan Jenkins1, Andrew Turner2, Mithun Krishnan3, Sachi Tripathi3, Sekhar Muddu4, Geet George3, S.S Angadi5

1Centre for Ecology & Hydrology, Wallingford, United Kingdom 2University of Reading, Reading, United Kingdom 3Indian Institue of Technology Kanpur, Kanpur, India 4Indian Institute of Science, Bangalore, India 5University of Agricultural Sciences, Dharwad, India

Small changes in precipitation and temperature can dramatically influence surface energy and water budgets in semi-arid regions. Quantifying land-atmosphere interactions and feedbacks in these areas is crucial to understanding global water and carbon cycles, for the development and testing of land surface, weather prediction and climate models, as well as for monitoring local water resources and agricultural output. We report the results of co-located observations of land surface water and energy fluxes and large-area soil moisture dynamics obtained at three study sites located across India. These sites were instrumented as part of the INCOMPASS (INteraction of Convective Organisation with Monsoon Precipitation, Atmosphere, Surface and Sea) and COSMOS-India projects. Two sites are located on contrasting red (Alfisols) and black (Vertisols) soils on the Deccan Plateau. A third site is installed on alluvial soils (Fluvisols) on the Indo-Gangetic Plain.  Each site consists of an eddy covariance flux tower providing measurements of sensible (H) and latent heat (LE) fluxes, micrometeorology and soil physics, in combination with a COSMOS (COsmic-ray Soil Moisture Observing System) sensor that provides spatially-integrated measurements of soil water content at field scale. In this presentation, we report on feedbacks between the land surface and the atmosphere, with a specific focus on the evaporative fraction (EF=LE/LE+H), precipitation and time varying soil moisture dynamics.

C407  -POSTER-0343: Designing a carbon and water observing system in wooded grasslands of Mpala and Ilmotiok sub-catchments in Laikipia, Kenya: in-situ observations based on a multi-scale approach

Stephen Kiama1, Richard Onwonga2, Geoffrey Kironchi2, Janeth Chepkemoi2

1Kenya Forestry Research Institute, Nairobi, Kenya 2University of Nairobi, Nairobi, Kenya

Rangelands are increasingly being degraded as a result of increased human and livestock populations. However, in Upper Ewaso Ngiro River Basin (UENRB) in Laikipia, Kenya, the impacts of these anthropogenic pressures are not well understood and their effects on carbon and water fluxes are also scarcely known. This is partly attributed to scarcity of information regarding their locations, spatial extent and degree of intensity. Deriving representations of these variables is complicated by surface heterogeneity as well as the effects of scale. The scale of investigation determines the range of patterns and processes identified, highlighting the need for appropriate scaling approach that resolves how landscape pattern and variability change with scale and allows proper representation of space. A central objective of this paper is to illuminate key aspects of surface heterogeneity that deserve consideration of scales and to highlight scaling approaches relevant to wooded grasslands. Our work is linked to an ongoing effort of establishing an observing system in Mpala and Ilmotiok sub-catchments in Laikipia. Intensive field campaigns conducted in the two sub-catchments integrate various ecological, hydrological, physiological, meteorological and climatological measurements sampled over several observational units oriented along the hydro-topographical gradients. Thus, we describe the design of this observing system as a separate objective, elaborating the components of the system and highlighting the significant scale-related issues and how they have been dealt with. For each (sub) component of the system, we have endeavoured to elaborate its provisions that would allow parameterization of biophysical (sub) models and scaling-up of measurements to landscape-level. Major assumptions guiding the scales of measurements defined for the unique observations are also highlighted. In general, the observing system is designed along a multi-scale approach, allowing integration of several variables that differs in terms of their scales of manifestation and capturing spatial and temporal variability of the landscape.

C408POSTER-0351: Role of local land - precipitation coupling in evolution of abnormally dry conditions in North Africa

Irina Petrova1, Diego Miralles1, Hendrik Wouters 1

1Ghent University, Ghent, Belgium

Droughts typically result in a severe social and economical damages in many parts of the world. Climate models predict intensification and proliferation of droughts in the future. Yet, knowledge of the mechanisms driving drought onset and development remains limited. Current understanding of the physics behind droughts suggests that persistent large-scale circulation anomalies are critical for their initiation, yet land-precipitation feedbacks may be central in their evolution. Understanding of physical processes behind this interaction is not yet well established.
 
In this study we use novel land surface and atmospheric satellite-based products and soil-water-atmosphere model to investigate the role of local land-atmosphere coupling in the evolution of abnormally dry conditions in North Africa. Specifically, we explore the difference in the impact of the soil moisture heterogeneity at 100 km scales on rainfall variability during very dry versus very wet years. Preliminary results suggest that during dry years, rain occurs more preferably over spatially drier soils and strong soil moisture gradients. The latter implies that abnormally dry conditions potentially increase effectiveness of the spatial soil moisture – precipitation coupling mechanism, which would operate to dampen established drought-like conditions. Further analyses will explore role of land surface state during the previous year on this year coupling relationship. Based on the obtained results a mechanism will be proposed to explain how land-precipitation coupling may influence intensity and evolution of droughts in the semi-arid climates.