Fine Time-Scale Winter Nutritional Ecology Of Deer Populations In The Lesser Khingan Mountains And Wandashan Mountais, Northeast China

Moose and roe deer fecal samples were collected during January from2009-2011in the Lesser Khingan Mountains, as well as red deer fecal samples in Wanda Mountains in Heilongjiang Province. We analyzed major food component of moose, roe deer and red deer by using detrench correspondence analysis. Then, we integrated the food component data in January to March2006and2007in the same research area together with data on2009-2011and established response surface model.In the mean time, we collected snow urine samples of moose and roe deer in January of2006,2007,2009and2010in the Lesser Khingan Mountains as well as red deer snow urine samples from Wanda Mountains in January and February of2009-2011. What’s more, we collected11pure urine samples from Alaska Moose Research Center and Chenney Mountain Zoo. In the field of conservation physiology, we chose urinary purine derivatives and their ratio to creatinine for the fine time-scale analysis, which could estimate deer nutritional condition preciously. We accurately calculated the daily digestible dry matter intake and metabolizable energy intake of red deer distributing in the Wanda Mountains and used a computational ecology method-BP artificial neural network to select and optimize the most influential climate factor to purine derivatives to creatinine ratios. Then response surface model was established using the selected climate data and purine derivatives to creatinine ratios as independent factors and response factors respectively. Then we enlarged the research time scale to years and discussed the influence of snow on purine derivatives. Generally, there are four major aspects of results:(1) In the Lesser Khingan Mountains, food diversity was the higest for moose in2009and roe deer in2011and in Wanda Mountains. According to the optimal forage theory, higher food diversity probability means that deers need more species to maintain their nutritional requirements, however, this statement could be biased based on the optimal foraging theory. Main food component of moose population distributing in the Lesser Khingan Mountains were poplar, willows, birch and spruce. Appearance of spruce showed a constrained nutritional condition in our research period. In contrast, the sympatric roe deer did not show a large component of needles in faces. In2009and2010main food component of roe deer were poplar, willows, birch, hazel and grass, while Mongolian oak became one of the main food components of roe deer in2011, and percentage of poplar, willows, birch became lower than2009and2010. This indicated that nutritional condition of roe deer might be better than moose in the Lesser Khingan Mountains. We found great difference between main food component in January and February of2009-2011and previous studies of red deer in Wanda Mountains. They do not select elm, amur linden and hazel as their main food but willows, birch and poplars. On one hand, this could be caused by different sampling time, on the other hand, the nutritional requirements changed under the influence of changing environmental condition.(2) During the establishment of response surface model, as we have already know that spruce has high content of tannins and that moose in the Lesser Khingan Mountains select spruce as one of their main food, so in order to avoid the influence of spruce, we tending to know that whether moose still prefer food with high tannin content without the influence of spruce. Moreover, we used the same model for the sympatric moose and roe deer to discover different response pattern of the two species. Modeling building results were significant.Moose distributing in the Lesser Khingan Mountains and red deer in the Wanda Mountains were tending to feed on food that have certain range of protein and tannins to reach a homostasis of their nutritional requirement, but this homostasis will fluctuate when environment changes. Generally, roe deer showed a preference of high protein and energy content food and avoidance of tannins, while still it is not difficult to figure out that when plant protein and tannins content achieve a certain range, roe deer showed a trend of preference to tannins.Results of purine derivatives detection in pure urine from free ranging mooses in North America indicated that, allantoin content was extremely low, and allantoin content in snow urine from wild population distributing in the Lesser Khingan Mountains was even lower than the detection limit of high performance chromatography. The low content of allantoin in the Lesser Khingan Mountains was supposed to be a strategy of antioxidant of adapting to extremely cold winter by this might be the same strategy to moose selected by roe deer in the Lesser Khingan Mountains to resist cold weather. But this still need more accurate experiment before we could confirm this hypothesis. And there is no allantoin absence phenomenon in red deer population in the Wanda Mountains, which might because the red deer population distributing in a low latitude and do not need this conservation strategy.During the sampling period in2006,2007,2009and2010, purine derivatives to creatinine ratios showed that both roe deer and moose populations were suffering from nutritional constrain, and the situation in2009is worse. This result was different from that we get from food component and food strategy analysis.According to the exist mathematical relationships between allantoin to creatinine ratio and digestible dry matter intake and metabolizable energy intake in female adult red deers, we trimmed off part of the allantoin to creatinine ratio from both sides of data distribution to keep the left data could best reflect the condition of adult females in the Wanda Mountains. Then we calculated the digestible dry matter intake and metabolizable energy intake of red deer by solving equations. Results showed that digestible dry matter intake was in short supply in both2009and2010. Metabolizable energy intake meet the minimum requirement of red deer in2010but in2009part of the individuals was short in metabolizable energy intake supply. These demonstrated that January-February in2009is a relative serious time and some supplementary (4) Climate factors did not show a common regulation, though there was some correlations among factors, but we could not judge which have the greatest influence. Therefore, we used artificial neural network, which is one of the computational ecology methods used for nonlinear analysis and modeling, to select and optimize the climate factors that had the most influential effects. Results showed that daily minimum temperature and daily average temperature had the best prediction effects to purine derivatives to creatinine ratios. Then response surface model were built to describe the mathematical relationships of daily minimum temperature and daily average temperature and urinary purine derivatives to creatinine ratio. Surprisingly, snow depth did not have good prediction effect. Then we enlarged our research time scale to years we found that snow depth should be used in long time scale nutritional ecology research.