|Title||Atmospheric effects of irrigation in monsoon climate: the Indian subcontinent|
|Source||University. Promotor(en): Pavel Kabat, co-promotor(en): Ronald Hutjes. - S.l. : s.n. - ISBN 9789461734990 - 166|
Earth System Science
Alterra - Climate change and adaptive land and water management
|Publication type||Dissertation, internally prepared|
|Keyword(s)||irrigatie - atmosfeer - moesson - evaporatie - waterbudget - india - irrigation - atmosphere - monsoon - evaporation - water budget|
During the 20th century, an increasing population increased the demand for food. As a consequence, agricultural activity has expanded and become more intense. A part of this intensification is the use of irrigation systems to water crops. Due to this irrigation, dams and channeling systems, water can be made available for agriculture in places or during seasons with limited precipitation. In monsoon climates, such as India, the majority of the precipitation falls in one season. During the rest of the year, water that is stored in dammed reservoirs can be made available to spread the water availability more evenly over the year. Previous studies with large scale hydrological models have shown that as a consequence of human influences (such as dams and irrigation systems), the river flow decreases during the wet monsoon months, but the evaporation of water into the atmosphere increases during the dry months. However, these large scale hydrological models did not take into account the atmospheric effects of a changed land surface. This PhD research studies these atmospheric effects of large scale irrigation in India. Three perspectives are taken to determined the influence of irrigation: (1) the local effects of a moister land-surface on the triggering of precipitation (i.e. does the change in land surface wetness lead to a different amount of precipitation?), (2) the atmospheric fate of evaporation due to irrigation (i.e. where does the evaporation lead to (down-wind) precipitation?), and (3) the effects of a moister land-surface on the large scale (monsoon) moisture transport patterns (i.e. do the monsoon flows change significantly due to large scale irrigation?) In the first part (the first perspective), several land-atmosphere diagnostics are tested globally. The goal of these diagnostics is to determine the influence of the land surface on precipitation, based on surface and atmospheric conditions. Of these diagnostics, the CTP-HIlow framework (Convective Triggering Potential and Humidity Index of the LOWer atmosphere) of Findell and Eltahir (2003a) performed well globally and over the Indian region. The summertime atmospheric conditions were diagnosed using this framework and the presence of a land-atmosphere coupling hot-spot in the Indian peninsula, proposed by previous studies (Koster et al. (2004)), is confirmed. Secondly, the local perspective is taken in the Indian subcontinent. The CTP- HIlow framework is tested in India, using an atmospheric slab model (a simple, one- dimensional model of the atmosphere) combined with atmospheric soundings (balloon measurements of temperature and moisture of an atmospheric profile of up to 30 km). This model is run twice; once with a wet land surface and once with a dry land surface. The results of these model runs can have two outcomes; the land surface does not have an influence on precipitation or it does have an influence. The CTP-HIlow framework proves to be useful to classify the potential influence of the land surface. When the atmosphere is very wet (low values of HIlow ), precipitation will occur regardless of the land surface, when the atmosphere is very dry (high HIlow values) no precipitation will occur, regardless of the land surface. However, for intermediate HIlow values, the effect of the land surface depends on the stability or the amount of convective energy (CTP) in the atmosphere. The stability of the atmosphere is related to how fast a particle will ascend in the atmosphere, which depends mostly on the temperature profile. For positive, but low convective potentials (0<CTP<200 J/kg), a wet land surface will produce more precipitation than a dry land surface. However, for high amounts of convective potential (CTP>200 J/kg), a dry land surface will produce more precipitation. For India, a small adaptation of the framework improved the performance in predicting the influence of the land surface on precipitation triggering. For India, the effect of the land surface on precipitation is seasonal. During the periods two months before the monsoon onset and after the monsoon retreat, precipitation triggering was found to be sensitive to land surface wetness. During those periods, a wet land surface is expected to increase precipitation. The atmospheric conditions under which a wet land surface is expected to decreases precipitation do not occur frequently in India. During the dry winter season, the atmosphere is too dry for the land surface to have an influence on precipitation. During the monsoon period, the atmosphere is too wet for the land surface to have an influence on precipitation, it will occur regardless of the land surface conditions. In the third part of the study, the moisture recycling perspective was taken and the atmospheric moisture budget of the Ganges basin is studied. A three-dimensional moisture tracing model is used to release moisture parcels from the Ganges basin, similarly to a class of school children releasing helium-filled balloons with their address on it. These parcels were transported along the wind patterns. During the trajectory of the parcel through the atmosphere, some moisture will precipitate out of it and contribute to the precipitation at that location. For each location, many parcels were released for every time step of 6 hours. Similar to the balloons of the school children that are hopefully sent back to them, the fate of the released moisture was accounted. The fraction of the evaporated moisture that subsequently falls as precipitation (recycles) within the Ganges basin shows a strong seasonality. During the winter months, practically all evaporation parcels were transported towards the Indian ocean and were lost for the Ganges basin. During the pre-monsoon months, the recycled fraction increased and was between 30-40%. During the monsoon months, the recycling peaks at up to 60%, after which it drops off again. The importance of recycled evaporation to the total precipitation peaks during the pre-monsoon and post-monsoon periods, when it contributes up to 15% of the precipitation. In the last part of the study, the effects from the local and moisture recycling perspectives are compared to those from the large scale perspective. Four atmospheric models were run with and without irrigation to test the large scale effects of irrigation on the Ganges basin atmospheric water budget and the influence on large scale atmospheric moisture transport. The local effects on precipitation were minimal and not uniform across the models. The Ganges river basin evaporation increased, as well as the amount of evaporation recycled within the river basin. However, the large scale wind patterns showed an uniform change across the models. Due to an increased flow in the direction of north-west India, the precipitation in east-India decreased while it increased in north-west India and Pakistan. Therefore, the Ganges basin precipitation decreased slightly. The conclusion of the work is that from the local perspective and the moisture recycling perspective, irrigation will lead to more precipitation in India. A wetter land surface will trigger some additional precipitation (especially just before and after the monsoon season) and a significant fraction of the evaporation will return to the same river basin as precipitation. However, from the large scale perspective, large scale irrigation will shift the wind patterns due to changes in the land-sea temperature contrast; precipitation will decrease slightly in the Ganges basin and be shifted towards the Indus basin and north-west India. The effects of irrigation on precipitation is small compared to the hydrological response of human influences simulated by the large scale hydrological models. Moreover, the spread in response across these hydrological models is large compared to the simulated effects of irrigation by the atmospheric models. Therefore, it is recommended to improve the large scale hydrological models and reduce their uncertainty before including the feedbacks of land use changes on their precipitation input.