Assessment of the transpirative deficit index for the Lombardy plain (Northern Italy)

Water scarcity describes a situation of recurrent water imbalance, where water demand exceeds natural renewable availability, on a timescale of months, seasons or years. A recent report by the European Union shows that, due to the effects of global (socio-economic and climate) changes on water resources, water scarcity shall affect by 2030 many areas of Europe characterized by large water availability but also by a high water consumption, such as the Po River Plain (Northern Italy). In the Po River Plain, irrigated crops cover more than 70% of the agricultural land, surface irrigation methods are largely applied, and massive amounts of water are diverted from rivers for irrigation. Additionally, an intensive agriculture coexists with highly developed industrial and commercial activities and with a considerable population, all requiring significant water volumes. Nowadays, the Po River Plain is not a water scarce basin, but short-term water shortages (i.e. on a timescale of days or weeks) can occasionally occur in some areas, especially in years of drought (i.e. temporary, negative and severe deviations from average precipitation values on a mid-term timescale) such as 2003. The Lombardy Plain covers a quarter of the Po River Plain, and represents the study area for this work. The European Drought Observatory (EDO) focuses its efforts to drought predictability, by using statistical analysis of historical occurrences and numerical weather forecasts. The EDO working group regularly generates maps of drought indicators through all the EU, using both remote sensing and hydrological modelling. Simulations with hydrological models are always conducted in absence of irrigation inputs. However, a comprehensive analysis of water scarcity, which can be useful to agricultural water managers and farmers, requires a joint assessment of information about water availability and consumption. This analysis can not be conducted at the European scale, both for the coarseness of the hydrological modelling simulation conducted by EDO, and for the peculiarity of the Po River Plain irrigation system in which irrigation inputs provided by many irrigation water sources play a fundamental role in the soil water balance. In this context, a study was start to develop a set of indicators for Water Scarcity and Drought (WS&D) suitable for the assessment of the state of irrigated agriculture in Northern Italy, to be used by regional policy makers, agricultural water managers and farmers. Indicators will be based on the synergic use of hydrological modelling and earth observation information applied at a spatial scale of interest for end-users (i.e. cells of 250 m). As a first step, this work presents the implementation and the application to the entire Lombardy Plain of an agricultural drought index called Transpirative Deficit Index (DTx). DTx is based on the transpiration deficit (calculated as the difference between potential and actual transpiration) computed daily by a spatially distributed water balance model (IDRAGRA) and cumulated over a period of x days (where x is 10, 30, 60 or 90). We firstly computed this index in absence of irrigation inputs. The simulation model, IDRAGRA, is a distributed-parameters conceptual model, developed by the Section of Agricultural Hydraulics of the Department of Agricultural and Environmental Sciences (DiSAA) of the University of Milan, which allows the simulation of the irrigation water distribution and the computation of the hydrologic balance on a daily basis. The model core is a soil-crop water balance module, which accounts for spatial variability of soils, crops, meteorological and irrigation inputs by dividing the study area with a regular mesh (i.e. a 250 m resolution grid). Each cell identifies a soil volume, subdivided into two layers, modelled as non-linear reservoirs in cascade, where hydrological processes are represented as one-dimensional. Moreover, IDRAGRA includes modules devoted to the simulation of crop development as a function of air temperature, and of water sources conveyance and distribution over the territory. We applied the model to the Lombardy Plain for the temporal horizon 1993-2007 (15 years), calculating DTx indices for each x value. We compared DTx series for each day of a specific year to the corresponding DTx values of the historical series. In particular, parameters of the normal distribution fitting the 15 DTx values were calculated, and the position of DTx for each year was obtained in term of probability of exceeding a threshold percentile of the distribution. In this work we present the results for two years, respectively characterized by drought (year 2003) and abundant rainfall (year 2008) throughout Northern Italy. The DTx was computed conducting three simulations, each one for a hypothetical uniform land use: a reference crop (i.e. a hypothetical crop resembling an extensive surface of green grass of uniform height, actively growing and adequately watered), maize and permanent grass (these last two crops are very widespread in the Lombardy region). We supposed each crop uniformly distributed over the study area and considered and, for each simulation, only water supplied by rainfall (i.e. absence of irrigation). In that way, the distribution of DTx values reflects the potential water stress conditions for that crop linked to a specific pedo-climatic situation. For all the crops, the distribution of DTx reflects the intra-annual differences in crop growth during the vegetative season, based on the meteorological and soil variability. In May, maize do not emerge at the same time throughout the study area, and until June DTx values are usually lower than those of the reference crop. In July, as crops are completely grown, DTx become higher than the reference. In conclusion, DTx can be a useful tool in the monitoring of the agricultural drought at the regional scale, reflecting the pedo-climatic variability throughout the study area. In the next future, we are planning to include the effective land use and the irrigation supply into the hydrological simulation, to evaluate the suitability of the DTx also for water shortage assessment.